1
00:00:00,200 --> 00:00:02,500
Mental will be is a silent 
pandemic. 

2
00:00:02,900 --> 00:00:06,800
According to the who depression 
and anxiety caused the global 

3
00:00:06,800 --> 00:00:10,600
economy over one-trillion u.s. 
dollar every year, it's time to 

4
00:00:10,600 --> 00:00:13,600
make a difference. 
Learn how to enhance your life 

5
00:00:13,600 --> 00:00:15,800
through a master class from 
Founders well-being. 

6
00:00:15,800 --> 00:00:19,300
And my good friend, Sunday prowl
on mental Wellness, visit 

7
00:00:19,300 --> 00:00:23,900
Founders well-being.com / Master
Class to enroll and enter tlj 

8
00:00:24,000 --> 00:00:27,600
24. 
A 20% discount be a better 

9
00:00:27,600 --> 00:00:29,800
version of yourself and make an 
impact. 

10
00:00:30,100 --> 00:00:33,300
The ISA QB software architecture
Gathering is the International 

11
00:00:33,300 --> 00:00:36,200
Conference highlight for all 
those working on solution 

12
00:00:36,200 --> 00:00:39,900
structures in IT, projects 
primarily software Architects 

13
00:00:39,900 --> 00:00:43,500
developers and Professionals in 
quality assurance but also 

14
00:00:43,500 --> 00:00:46,400
system analysts who want to 
communicate better with their 

15
00:00:46,400 --> 00:00:48,800
developers. 
A selection of well-known 

16
00:00:48,800 --> 00:00:51,800
International speakers will 
share their practical knowledge 

17
00:00:51,800 --> 00:00:54,800
on the most important topics in 
the state of the art software 

18
00:00:54,800 --> 00:00:57,200
architecture. 
The conference takes place 

19
00:00:57,200 --> 00:01:02,200
online, from November, 14 to 17.
And we have a 15% discount code 

20
00:01:02,200 --> 00:01:05,400
for you. 
Enter tlj underscore MP 

21
00:01:05,400 --> 00:01:08,500
underscore 15 for 15 percent 
discount. 

22
00:01:08,800 --> 00:01:12,300
If they want to get value from 
their data, generally 

23
00:01:12,300 --> 00:01:17,400
data-driven value at the one to 
do that by applying analytics 

24
00:01:17,400 --> 00:01:21,600
and AI in almost every aspect of
their business. 

25
00:01:22,100 --> 00:01:26,100
They need to utilize data for 
both aspects, and all touch 

26
00:01:26,100 --> 00:01:29,700
points and all applications 
inside the company and outside. 

27
00:01:30,100 --> 00:01:34,000
If they have such a mission and 
they want to do that reliably 

28
00:01:34,000 --> 00:01:38,600
resilient lie and do that at 
scale fast, then they've got to 

29
00:01:38,600 --> 00:01:41,500
consider this dish. 
It's all about really are 

30
00:01:41,508 --> 00:01:49,000
locking the value of the data. 
Hey everyone. 

31
00:01:49,600 --> 00:01:51,400
My name is Henry Surya with 
Robin. 

32
00:01:53,100 --> 00:01:56,000
And you're listening to the 
technology, you know, podcast 

33
00:01:56,400 --> 00:01:58,700
the show where I'll be bringing 
you the greatest technical 

34
00:01:58,700 --> 00:02:02,500
leaders practitioners and 
thought leaders in the industry 

35
00:02:02,900 --> 00:02:07,300
to discuss about their Journey 
ideas and practices that we all 

36
00:02:07,300 --> 00:02:10,900
can learn and apply to build a 
highly performing technical team

37
00:02:11,300 --> 00:02:13,700
and to make an impact in your 
personal work. 

38
00:02:14,200 --> 00:02:22,400
So let's dive into our Journal. 
Hello again, my friends at my 

39
00:02:22,400 --> 00:02:24,200
listeners, welcome to the 
technology. 

40
00:02:24,200 --> 00:02:27,000
Now podcast the show where you 
can learn about technical 

41
00:02:27,000 --> 00:02:30,000
leadership and Excellence from 
my conversations with great 

42
00:02:30,000 --> 00:02:32,700
thought, leaders in the tech 
industry and you're listening to

43
00:02:32,700 --> 00:02:36,400
the episode number 107. 
If this is your first time 

44
00:02:36,400 --> 00:02:38,900
listening to technology, you 
know, subscribe and follow the 

45
00:02:38,900 --> 00:02:42,400
show on your podcast app and on 
LinkedIn, Twitter and Instagram.

46
00:02:42,700 --> 00:02:45,600
And if you'd like to support my 
journey, creating this podcast 

47
00:02:45,700 --> 00:02:49,400
subscriber, A patron at Tech. 
Did you know dot f / Patron? 

48
00:02:49,900 --> 00:02:52,600
My guest for today's episode is 
Jean meghani. 

49
00:02:53,100 --> 00:02:57,400
Jama is the founder of data mesh
Concept in 2018 and since then 

50
00:02:57,400 --> 00:03:00,500
has been evangelizing it to the 
wider industry, including 

51
00:03:00,500 --> 00:03:04,300
writing, her latest book, titled
data mesh in this episode. 

52
00:03:04,300 --> 00:03:07,300
We discussed in depth about the 
data mesh concept which is 

53
00:03:07,300 --> 00:03:10,800
starting to become an industry. 
Trend nowadays, we started our 

54
00:03:10,800 --> 00:03:13,700
conversation by discussing the 
current challenges working with 

55
00:03:13,700 --> 00:03:18,200
data such as the outdated data. 
Approach and why the current 

56
00:03:18,200 --> 00:03:22,200
data tools are still inadequate,
jean-marc then describe data 

57
00:03:22,200 --> 00:03:25,800
mesh and why organizations 
should adopt it to generate data

58
00:03:25,800 --> 00:03:29,600
driven values at skill 
jean-marc, then explain the four

59
00:03:29,600 --> 00:03:33,500
core principles of data mesh, 
which include domain ownership 

60
00:03:33,500 --> 00:03:37,000
data, is a product, the 
self-serve data platform, and 

61
00:03:37,000 --> 00:03:39,400
the Federated computational 
governance. 

62
00:03:40,100 --> 00:03:43,100
I really enjoyed my conversation
with jayamma learning the data 

63
00:03:43,100 --> 00:03:46,100
mesh Concept in depth, which has
been something, I would love to 

64
00:03:46,100 --> 00:03:48,700
learn more More about and this 
episode taught me a lot about 

65
00:03:48,700 --> 00:03:50,800
it. 
If you also enjoy listening to 

66
00:03:50,800 --> 00:03:53,400
this episode, will you help 
share it with your friends and 

67
00:03:53,400 --> 00:03:56,600
colleagues who can also benefit 
from listening to this episode? 

68
00:03:57,000 --> 00:04:00,300
My ultimate mission is to spread
this podcast to more listeners 

69
00:04:00,600 --> 00:04:03,500
and I really appreciate your 
support in any way towards 

70
00:04:03,500 --> 00:04:06,100
fulfilling my mission. 
Before we continue to the 

71
00:04:06,108 --> 00:04:09,400
conversations, let's hear some 
words from our sponsors 

72
00:04:09,700 --> 00:04:12,200
definitey is the top 
International software 

73
00:04:12,200 --> 00:04:16,200
development conference with an 
emphasis on coding architecture.

74
00:04:16,200 --> 00:04:19,399
And Leadership skills. 
The lineup for this year is 

75
00:04:19,399 --> 00:04:23,200
truly stellar and features many 
Legends in software development 

76
00:04:23,500 --> 00:04:25,700
names. 
Such as Robert Uncle Bob, Martin

77
00:04:26,000 --> 00:04:31,000
can back Scott Hanselman, Franca
subramanyam Carolyn honey, Alan 

78
00:04:31,000 --> 00:04:34,400
Halep, Mary, poppendieck, and 
many other prominent names, 

79
00:04:34,500 --> 00:04:37,700
including some of those who have
also appeared in this podcast 

80
00:04:37,700 --> 00:04:40,700
before the conference takes 
place online. 

81
00:04:40,800 --> 00:04:43,400
So you can enjoy it from the 
comfort of your couch. 

82
00:04:43,900 --> 00:04:46,900
We spoke to the definitey 
organizers and I'm happy to 

83
00:04:46,900 --> 00:04:49,400
share that technology. 
You know, has got the 10% 

84
00:04:49,400 --> 00:04:54,000
discount code for you. 
Enter the promo code, awsm 

85
00:04:54,200 --> 00:04:57,300
underscore tlj. 
When you purchase the ticket on 

86
00:04:57,300 --> 00:05:00,300
definite e.com, here's the promo
code. 

87
00:05:00,300 --> 00:05:04,400
One more time awsm underscore, 
tlj. 

88
00:05:05,100 --> 00:05:08,400
Depending on the time when you 
purchase a ticket early price is

89
00:05:08,400 --> 00:05:10,700
still available. 
See you there. 

90
00:05:11,100 --> 00:05:14,400
Today's episode is proudly 
sponsored by skills matter. 

91
00:05:14,700 --> 00:05:18,400
The global community and events.
Form with more than 100,000 

92
00:05:18,400 --> 00:05:22,200
software professionals here 
members, can organize their 

93
00:05:22,200 --> 00:05:24,900
learning experiences around the 
technology topics. 

94
00:05:24,900 --> 00:05:28,700
They care about most you get 
on-demand access to their latest

95
00:05:28,700 --> 00:05:32,000
content thought, leadership 
insights, as well as the 

96
00:05:32,000 --> 00:05:36,200
exciting schedule of tech events
running across all time zones. 

97
00:05:36,600 --> 00:05:40,200
So whether devops our data 
science is your bus or you are 

98
00:05:40,200 --> 00:05:44,100
fan of functional programming or
all things Cloud, you can make 

99
00:05:44,100 --> 00:05:47,800
real connections with people who
share your Our interests head on

100
00:05:47,800 --> 00:05:50,900
over to skills method of calm to
become part of the tech 

101
00:05:50,900 --> 00:05:53,000
community that matters most to 
you. 

102
00:05:53,300 --> 00:05:56,400
It's free to join and you will 
find it easy to keep up with the

103
00:05:56,407 --> 00:06:01,000
latest tech Trends. 
Hello everyone. 

104
00:06:01,000 --> 00:06:04,100
Welcome to another new episode 
of the package you know podcast 

105
00:06:04,200 --> 00:06:06,700
today I'm so excited to meet 
yarmulke ghani. 

106
00:06:06,900 --> 00:06:10,100
She was last a director of 
emerging Technologies in 

107
00:06:10,100 --> 00:06:12,900
thoughtworks. 
She was there probably around 11

108
00:06:12,900 --> 00:06:15,400
years, my nose. 
I'm not following her work when 

109
00:06:15,400 --> 00:06:17,000
I was working in thought. 
It's as well. 

110
00:06:17,200 --> 00:06:19,600
So she was part of the 
technology radar comedy and 

111
00:06:19,600 --> 00:06:22,400
always come up with all these 
emerging Technologies and 

112
00:06:22,400 --> 00:06:25,200
recently, in the last few years,
jean-marc came out with this 

113
00:06:25,200 --> 00:06:28,500
concept called Data mesh. 
I think it was around in 2018, 

114
00:06:28,500 --> 00:06:30,600
if I'm not wrong. 
Since then, the data mesh 

115
00:06:30,600 --> 00:06:34,100
concept has taken a surprise by 
many people and many people Rave

116
00:06:34,100 --> 00:06:36,500
about it. 
So today we'll be talking a lot 

117
00:06:36,500 --> 00:06:39,200
about data mesh and I'm really 
looking forward to have this 

118
00:06:39,200 --> 00:06:43,000
conversation with you Jama. 
It's a pleasure to be here and 

119
00:06:43,000 --> 00:06:45,900
really thank you for that 
perfect pronunciation of my 

120
00:06:45,900 --> 00:06:46,600
name. 
Mmmmm. 

121
00:06:47,200 --> 00:06:49,000
Oh, surprise that. 
Okay, yeah. 

122
00:06:50,800 --> 00:06:53,800
I would love to probably know 
more about your career. 

123
00:06:54,000 --> 00:06:57,500
So I always stop to ask my 
guests to share their career 

124
00:06:57,500 --> 00:07:00,900
Journey or any turning points or
highlights in their career maybe

125
00:07:00,900 --> 00:07:02,800
if you can share a little bit 
about yourself. 

126
00:07:03,300 --> 00:07:06,300
Sure. 
Laughter I guess my journey is 

127
00:07:06,300 --> 00:07:11,200
filled with detours and going to
new places led by curiosity. 

128
00:07:11,200 --> 00:07:15,000
So I started as a software 
engineer for the first 14 years 

129
00:07:15,000 --> 00:07:18,400
of my career. 
I worked in Take RND companies, 

130
00:07:18,400 --> 00:07:21,300
where they were building a 
technology product side. 

131
00:07:21,300 --> 00:07:25,600
Dates distributed systems before
cloud and large-scaled. 

132
00:07:25,600 --> 00:07:27,800
It's really systems. 
Was thing. 

133
00:07:28,100 --> 00:07:32,200
Build monitoring and 
observability from scratch 

134
00:07:32,200 --> 00:07:35,200
building steering systems, 
building databases that 

135
00:07:35,200 --> 00:07:38,200
basically gets signals from 
critical infrastructure. 

136
00:07:38,200 --> 00:07:42,000
Analyze those signals turning 
them into reports and so on, the

137
00:07:42,000 --> 00:07:46,500
full stack of what real world, 
this was so could look like And 

138
00:07:46,500 --> 00:07:49,400
I did that on multiple operating
systems, various laser is a 

139
00:07:49,400 --> 00:07:52,200
feelings and HP Tandem and 
windows and so on, that really 

140
00:07:52,200 --> 00:07:56,400
gave me a good Insight, bottom 
up to the technologies. 

141
00:07:56,400 --> 00:07:59,300
That I realize that a lot of 
Technologies today. 

142
00:07:59,300 --> 00:08:02,400
Maybe start with web development
or app development, and they 

143
00:08:02,400 --> 00:08:05,600
don't get the opportunity to 
really like look inside the 

144
00:08:05,600 --> 00:08:07,100
kernel. 
And how to do system 

145
00:08:07,100 --> 00:08:08,800
programming, how to build 
product calls. 

146
00:08:09,000 --> 00:08:10,700
I did all of that, which was 
awesome. 

147
00:08:11,200 --> 00:08:13,400
And then, I took a bit of a 
detour. 

148
00:08:13,700 --> 00:08:15,500
It went to a hardware for a 
little bit. 

149
00:08:15,500 --> 00:08:18,300
I work for a company. 
It was building various Hardware

150
00:08:18,300 --> 00:08:20,800
from scratch. 
We were building the firmware on

151
00:08:20,800 --> 00:08:24,700
digital pen systems and that was
interesting as well. 

152
00:08:24,900 --> 00:08:29,000
Again, a lot of great learnings,
how to build embedded firmware 

153
00:08:29,400 --> 00:08:32,200
and then, another detour I came 
to Consulting without works. 

154
00:08:32,400 --> 00:08:35,200
That's we worked at the worst. 
We work on have largest skill. 

155
00:08:35,200 --> 00:08:38,900
Exact you ssion. 
So that led to microservices and

156
00:08:38,900 --> 00:08:41,600
again building large scale. 
I guess this resolutions with 

157
00:08:41,600 --> 00:08:44,300
micro services and that 
ecosystem like did quite a bit 

158
00:08:44,300 --> 00:08:47,100
of work in that space for quite 
a while and I was Excited about 

159
00:08:47,100 --> 00:08:50,100
service mesh and kubernetes and 
all of the Technologies. 

160
00:08:50,100 --> 00:08:53,800
That's really made it possible. 
And as you said about four years

161
00:08:53,800 --> 00:08:57,300
ago, I kind of started putting 
my nose into the data space to 

162
00:08:57,300 --> 00:09:02,600
my surprise, the world of data 
was far from the agility and 

163
00:09:02,600 --> 00:09:06,100
nimbleness and totally and 
distribution decentralization, 

164
00:09:06,100 --> 00:09:09,000
that was seen active world. 
It's slow. 

165
00:09:09,000 --> 00:09:13,600
Moving is based on paradigms 
around centralization of data 

166
00:09:13,600 --> 00:09:16,100
and middleman moving data around
and up. 

167
00:09:16,400 --> 00:09:19,100
It was very sad. 
Observation, to be honest. 

168
00:09:19,100 --> 00:09:22,600
That I thought, okay, I could be
the little kid who points the 

169
00:09:22,600 --> 00:09:25,400
bit naked Emperor. 
And I don't like doing that even

170
00:09:25,400 --> 00:09:28,400
if I get attacked. 
So I started like talking about 

171
00:09:28,400 --> 00:09:30,100
capping would shift the parents 
on. 

172
00:09:30,300 --> 00:09:33,000
What are the pain points? 
Like wake up people? 

173
00:09:33,400 --> 00:09:33,800
I am. 
Yeah. 

174
00:09:33,800 --> 00:09:35,600
And they suggest really came as 
if I caused. 

175
00:09:35,600 --> 00:09:40,600
This is a question with some 
answers, some basic answers 

176
00:09:40,900 --> 00:09:44,000
since then I've been kind of 
building it, evangelizing it. 

177
00:09:44,000 --> 00:09:47,300
And as of two weeks ago, I 
decided to start A company to 

178
00:09:47,300 --> 00:09:51,800
build a product, a technology 
deep take, kind of Dev product, 

179
00:09:51,800 --> 00:09:54,400
developer facing products that 
it's going to make it. 

180
00:09:54,400 --> 00:09:59,500
So easy to kind of work with 
data under the mesh principles 

181
00:09:59,900 --> 00:10:03,200
to really show the world a 
different way of doing things 

182
00:10:03,300 --> 00:10:06,800
and enable developers. 
Most importantly, while sounds 

183
00:10:06,800 --> 00:10:07,700
real, exciting. 
Yeah. 

184
00:10:07,700 --> 00:10:09,800
I saw your post about quitting 
Pollock's job. 

185
00:10:09,800 --> 00:10:13,100
It was surprising to me but I 
think looking at the opportunity

186
00:10:13,100 --> 00:10:16,300
of investing more effort in 
building the tools about data. 

187
00:10:16,300 --> 00:10:19,400
I think that will be definitely 
crucial and will help a lot of 

188
00:10:19,408 --> 00:10:21,500
people. 
But before we go into this 

189
00:10:21,500 --> 00:10:24,000
exciting journey, I think. 
First of all, the topic of this 

190
00:10:24,000 --> 00:10:26,800
conversation is about data mesh.
And you wrote a book with the 

191
00:10:26,800 --> 00:10:29,800
same title data mesh with the 
subtitle of delivering data 

192
00:10:29,800 --> 00:10:33,500
driven value at scale. 
So maybe if you can share a 

193
00:10:33,508 --> 00:10:37,400
little bit, what did you 
actually see the problems when 

194
00:10:37,400 --> 00:10:41,000
you were working with the data? 
You mention about some old ways 

195
00:10:41,000 --> 00:10:43,100
Paradigm centralization and 
things like that. 

196
00:10:43,100 --> 00:10:46,200
But what kind of challenges and 
problems when you work? 

197
00:10:46,300 --> 00:10:49,300
With the data problem during 
that time, and maybe if you can 

198
00:10:49,300 --> 00:10:53,400
give an overview, how data has 
always been approached in the 

199
00:10:53,400 --> 00:10:56,200
delivery or in the day-to-day 
development. 

200
00:10:56,600 --> 00:10:59,500
Yeah, that's so good. 
Way of positioning data mesh. 

201
00:10:59,900 --> 00:11:04,800
The problem that I saw was that 
we had an assumption that to get

202
00:11:04,800 --> 00:11:09,100
value from data and value being 
through Pi business intelligence

203
00:11:09,100 --> 00:11:11,700
through reporting, through 
training machine, learning 

204
00:11:11,700 --> 00:11:16,200
models, or all sorts of analysis
of data data needs to come on. 

205
00:11:16,400 --> 00:11:21,000
It's modeled or raw and skit. 
Centralized that was a response 

206
00:11:21,000 --> 00:11:24,400
to a siloing of data, in 
applications databases. 

207
00:11:24,400 --> 00:11:27,400
That really didn't allow 
cross-cutting analysis of data 

208
00:11:27,400 --> 00:11:31,200
across town occasions, but that 
centralization that idea of 

209
00:11:31,200 --> 00:11:36,100
centralization of had led to a 
very fragile, very slow-moving, 

210
00:11:36,200 --> 00:11:39,800
architecture and bottlenecks for
really getting value from data. 

211
00:11:39,800 --> 00:11:42,200
So what are the points of 
fragility? 

212
00:11:42,200 --> 00:11:46,200
That causes so much waste? 
Those are data Pipelines. 

213
00:11:46,600 --> 00:11:49,100
So to centralized data from 
applications, we have this 

214
00:11:49,100 --> 00:11:53,100
concept of ETL elt, you know, 
extraction of data for 

215
00:11:53,100 --> 00:11:55,500
application databases. 
I think it's curved the biggest 

216
00:11:55,500 --> 00:11:58,200
clients and we can do it. 
Architecture is so intrusive, 

217
00:11:58,200 --> 00:12:00,000
right? 
Because there's no contract. 

218
00:12:00,000 --> 00:12:04,100
There is no abstraction. 
So you have constantly breaking 

219
00:12:04,200 --> 00:12:07,600
very task-oriented job. 
Oriented a labyrinth pipelines 

220
00:12:07,600 --> 00:12:11,500
are complex fireflies, moving 
stuff around and converting them

221
00:12:11,600 --> 00:12:14,000
putting the from one thing to 
another thing. 

222
00:12:14,300 --> 00:12:17,900
And that is a very fragile and 
Causes a lot of waste. 

223
00:12:18,400 --> 00:12:21,300
By the time the data has popped 
out of the other end of the 

224
00:12:21,300 --> 00:12:25,200
pipeline, The Source has moved 
on and you got problems to solve

225
00:12:25,700 --> 00:12:27,800
and then on the other hand you 
have this big bottleneck. 

226
00:12:27,800 --> 00:12:31,100
So the assumption that there is 
a data team responsible for the 

227
00:12:31,100 --> 00:12:34,000
data from everywhere, and they 
put it in a warehouse of Lake. 

228
00:12:34,100 --> 00:12:37,100
It is a flawed assumption, in an
organization that needs to move 

229
00:12:37,100 --> 00:12:41,900
fast, leads to share data. 
More peer-to-peer, they become a

230
00:12:41,900 --> 00:12:44,000
bottleneck, the team of the 
architecture itself, become a 

231
00:12:44,008 --> 00:12:46,400
bottleneck. 
So you have frustrated users As 

232
00:12:46,400 --> 00:12:48,600
data users, they can't find the 
data they need. 

233
00:12:48,600 --> 00:12:51,700
They can't access it. 
They don't trust it by the time 

234
00:12:51,700 --> 00:12:53,500
they to God made available to 
them. 

235
00:12:53,500 --> 00:12:56,800
The source has moved on data 
people in the middle that 

236
00:12:56,800 --> 00:12:59,700
honestly I have all the oven 
empathy for them in the world 

237
00:12:59,700 --> 00:13:02,800
that they have been. 
Given this impossible task of 

238
00:13:02,800 --> 00:13:05,400
getting data from people that 
have no intention of sharing 

239
00:13:05,400 --> 00:13:06,700
data. 
And giving it to people that 

240
00:13:06,700 --> 00:13:08,800
they have no idea how they're 
going to use it and it's stuck 

241
00:13:08,800 --> 00:13:12,100
in the middle like waiting. 
They do watching and troubling 

242
00:13:12,100 --> 00:13:15,200
data with no ready purpose to be
honest. 

243
00:13:15,500 --> 00:13:19,200
They are under a The pressure 
and the don't perform and then 

244
00:13:19,200 --> 00:13:22,900
the applications really don't 
have much of a visibility or 

245
00:13:22,900 --> 00:13:24,900
even opportunity to use that 
data. 

246
00:13:25,100 --> 00:13:27,800
They just put data how to put 
that in the databases and use it

247
00:13:27,800 --> 00:13:30,900
for their operational needs and 
they never in the conversation 

248
00:13:30,900 --> 00:13:34,100
in analytics. 
So they're isolated from real 

249
00:13:34,100 --> 00:13:37,400
application of possibility of 
the mail embedded into their 

250
00:13:37,400 --> 00:13:40,300
applications, but becoming 
analytical data users because if

251
00:13:40,300 --> 00:13:43,200
they've been got, just put a 
satellite from this data world. 

252
00:13:43,500 --> 00:13:44,800
Yeah. 
So fragile Leti. 

253
00:13:45,000 --> 00:13:49,600
The only time from day Attitude 
value, bottom legs, out of that 

254
00:13:49,600 --> 00:13:52,100
are major problems that you see 
with the past part. 

255
00:13:52,100 --> 00:13:56,400
And I think in my career so far,
I've experienced things like for

256
00:13:56,400 --> 00:14:00,100
example, big giant database 
where you have when the database

257
00:14:00,100 --> 00:14:03,500
like Oracle SQL Server, right? 
Then we move into the paradigms 

258
00:14:03,500 --> 00:14:06,500
of data warehouse concept, where
you have another set of tools. 

259
00:14:06,500 --> 00:14:09,500
So you pump the data, from the 
oil, TP database into this, 

260
00:14:09,500 --> 00:14:12,600
analytical data base. 
And then in the last, maybe 10 

261
00:14:12,600 --> 00:14:15,700
years we moved to another 
concept called Data Lake where 

262
00:14:15,700 --> 00:14:18,800
you probably Release raw data, 
put it there and from there you 

263
00:14:18,800 --> 00:14:22,100
move into different data Maps, 
may be of different small data 

264
00:14:22,100 --> 00:14:25,800
warehouse and now we move into 
this Cloud Model where we also 

265
00:14:25,800 --> 00:14:29,000
see Cloud Technologies. 
Helping I'm familiar much more 

266
00:14:29,000 --> 00:14:32,000
with like bigquery and Amazon, 
maybe you have red shift. 

267
00:14:32,000 --> 00:14:34,500
So if you see all these 
historical, the unique thing 

268
00:14:34,500 --> 00:14:36,800
about it, like you mentioned the
centralization, why people 

269
00:14:36,800 --> 00:14:40,000
actually move more towards 
centralization rather than the 

270
00:14:40,000 --> 00:14:43,500
decentralization part? 
Yes, easy to get started with 

271
00:14:43,500 --> 00:14:46,100
right when your problem space is
new. 

272
00:14:46,300 --> 00:14:50,200
New or solution space is new and
you require specialization. 

273
00:14:50,400 --> 00:14:53,700
You've got a new set of tools 
and new ways of working with 

274
00:14:53,700 --> 00:14:57,200
data. 
You can leverage your majority 

275
00:14:57,200 --> 00:15:01,400
of organization can really push 
the responsibility into every 

276
00:15:01,400 --> 00:15:03,100
team. 
You have to centralize a 

277
00:15:03,100 --> 00:15:08,200
specialized people and also it's
easier to control and easier to 

278
00:15:08,200 --> 00:15:10,600
get started with like we know 
right? 

279
00:15:10,700 --> 00:15:13,400
Any startup start building an 
application or solution. 

280
00:15:13,400 --> 00:15:16,100
We say like start with the money
list, find your Market. 

281
00:15:16,300 --> 00:15:18,400
It and then break it down 
because decentralized 

282
00:15:18,400 --> 00:15:21,300
distributed systems are 
inherently complex. 

283
00:15:21,700 --> 00:15:25,900
So there has to be a pivotal 
point that the complexity of 

284
00:15:25,900 --> 00:15:29,300
your environment increases to 
the point that centralized 

285
00:15:29,300 --> 00:15:33,800
Simple Solution does not work 
anymore and I think the data 

286
00:15:33,800 --> 00:15:37,700
warehouses or lakes or lake 
houses as an architectural 

287
00:15:37,700 --> 00:15:40,400
Paradigm, not so much of an 
underlying technology. 

288
00:15:40,700 --> 00:15:43,000
They've been suitable for this 
world. 

289
00:15:43,000 --> 00:15:46,100
That the data wasn't ubiquitous 
perhaps we were capturing. 

290
00:15:46,200 --> 00:15:47,700
And it'll from every touch 
point. 

291
00:15:47,700 --> 00:15:50,900
The data wasn't being used in 
every single application domain 

292
00:15:51,200 --> 00:15:53,600
and having a centralized team 
and decentralized Technology 

293
00:15:53,600 --> 00:15:55,600
stack to do with it was 
acceptable. 

294
00:15:55,900 --> 00:15:57,600
But we don't no longer needing 
that world. 

295
00:15:57,600 --> 00:16:00,300
We've passed that pivotal point 
of complexity. 

296
00:16:01,000 --> 00:16:03,800
So you mentioned about this 
point about operational forces, 

297
00:16:03,800 --> 00:16:07,400
analogy goal, divide where the 
application team probably just 

298
00:16:07,400 --> 00:16:10,200
dump data in the database. 
And maybe there's another set of

299
00:16:10,200 --> 00:16:13,200
data analyst data engineer, 
trying to get the data from 

300
00:16:13,200 --> 00:16:16,100
those databases and put it in a 
central place and getting in. 

301
00:16:16,300 --> 00:16:18,000
Sites. 
There are a lot of complexities 

302
00:16:18,000 --> 00:16:20,700
definitely but I think in the 
last few years we also see a lot

303
00:16:20,700 --> 00:16:22,800
of advancement in data 
Technologies. 

304
00:16:23,000 --> 00:16:25,300
So I think I saw in your book 
and presentation. 

305
00:16:25,300 --> 00:16:28,300
You have this one slide we're 
probably there are so many 

306
00:16:28,300 --> 00:16:31,500
different Technologies. 
The logos are too small, why 

307
00:16:31,500 --> 00:16:35,500
those tools still couldn't solve
this kind of problem because it 

308
00:16:35,500 --> 00:16:37,600
seems like there are so many 
advancement. 

309
00:16:38,200 --> 00:16:39,700
Yeah, that's a very good 
question. 

310
00:16:40,000 --> 00:16:42,800
I mean, the tools are solving 
these problems. 

311
00:16:43,000 --> 00:16:45,100
Oh, actually, collide with an 
example. 

312
00:16:45,300 --> 00:16:50,800
The nevertheless Yes, they are 
created for an operating model 

313
00:16:51,000 --> 00:16:54,300
that at the end of the day is 
pipeline data transform. 

314
00:16:54,500 --> 00:16:59,200
Put into storage layer with 
metadata, or spring, 40ml on top

315
00:16:59,200 --> 00:17:01,800
and voila. 
You get value on the other end. 

316
00:17:02,000 --> 00:17:04,900
So they have been organized 
around this very centralized 

317
00:17:04,900 --> 00:17:08,099
pipeline model at a very macro 
level, right? 

318
00:17:08,400 --> 00:17:13,300
If he's ooh Weld and if you 
start sprinkling, these tools to

319
00:17:13,300 --> 00:17:16,099
an overall kind of big picture 
meta. 

320
00:17:16,300 --> 00:17:18,599
Picture, that's what they're 
organized to do with it. 

321
00:17:18,599 --> 00:17:21,700
Organized to solve ingestion 
problems, that were organized to

322
00:17:21,700 --> 00:17:25,900
self big hairy part, line 
problems their flow. 

323
00:17:26,000 --> 00:17:28,700
Like those sort of Technology. 
They designed to solve the big 

324
00:17:28,700 --> 00:17:31,100
data storage or parallel 
processing problem. 

325
00:17:31,100 --> 00:17:35,100
So unless you change the 
operating model and that meta 

326
00:17:35,100 --> 00:17:39,400
architecture, no matter how 
locally you optimize a 

327
00:17:39,408 --> 00:17:43,400
particular solution, we're going
to optimize and solve the really

328
00:17:43,400 --> 00:17:45,100
hairy centralized data 
pipelines. 

329
00:17:45,300 --> 00:17:48,000
You still going to have a Harry 
satellite data pipeline, you get

330
00:17:48,000 --> 00:17:50,300
a little bit better and maybe 
detecting the errors of 

331
00:17:50,300 --> 00:17:53,800
connecting to more resources, 
But ultimately you're still 

332
00:17:53,800 --> 00:17:58,600
stuck in the past Paradigm and 
those fundamental assumptions 

333
00:17:58,600 --> 00:18:01,900
that need to be invalidated and 
need to change remain. 

334
00:18:02,400 --> 00:18:05,800
In fact, people who are deeply 
in the data space and try to use

335
00:18:05,800 --> 00:18:08,700
these tools, they say they 
struggle and they suffer. 

336
00:18:08,700 --> 00:18:11,500
They've got this massive 
landscape of fragmented 

337
00:18:11,500 --> 00:18:15,000
technologies that frankly work, 
really, with difficulty with 

338
00:18:15,000 --> 00:18:17,200
each other. 
The work with them have to 

339
00:18:17,300 --> 00:18:19,700
integrate themselves. 
That costs of integration looks 

340
00:18:19,700 --> 00:18:21,600
of this technology to a 
meaningful. 

341
00:18:21,600 --> 00:18:25,300
Scalable solution is very high. 
I mean, if you look at every 

342
00:18:25,300 --> 00:18:29,000
vendor on that diagram, and if 
you go to their connector Pages,

343
00:18:29,200 --> 00:18:32,500
this business is built around, 
just custom proprietary 

344
00:18:32,500 --> 00:18:35,000
connectors to yet. 
Another day's associate allowed,

345
00:18:35,000 --> 00:18:38,000
a distinct lack of 
standardization is just 

346
00:18:38,000 --> 00:18:41,400
mind-boggling in this space. 
So it's like the Tower of Babel,

347
00:18:41,400 --> 00:18:44,400
or like it's just falling apart 
in my mind, it just feels like 

348
00:18:44,400 --> 00:18:48,200
nobody speaks the same language.
And to a large degree tools are 

349
00:18:48,200 --> 00:18:51,700
built to solve a very custom 
solution, and the foundation of 

350
00:18:51,700 --> 00:18:53,500
this house is about to fall out 
of. 

351
00:18:53,700 --> 00:18:55,400
So, you still have a rocket 
Foundation. 

352
00:18:55,600 --> 00:18:59,000
I know, I can be very polarising
as I describe this because, I'm 

353
00:18:59,000 --> 00:19:02,200
very passionate about. 
Let's fix the foundation. 

354
00:19:02,200 --> 00:19:04,100
Let's rethink. 
Our operating model. 

355
00:19:04,900 --> 00:19:06,800
Totally makes sense. 
Because he had the way you 

356
00:19:06,800 --> 00:19:08,700
mentioned about cost of 
integration. 

357
00:19:08,800 --> 00:19:11,500
I can imagine every data 
products that I see, you will 

358
00:19:11,500 --> 00:19:15,300
see all these Integrations, then
the more the better after you do

359
00:19:15,300 --> 00:19:18,300
it actually the in Essman the 
kind of like lockdown and you 

360
00:19:18,300 --> 00:19:21,300
put so much effort and maybe 
money to actually move your 

361
00:19:21,300 --> 00:19:23,100
data. 
But eventually if you need to 

362
00:19:23,100 --> 00:19:26,300
switch also the costs for you to
move the data out to the 

363
00:19:26,300 --> 00:19:28,700
different Technologies. 
I think that's so painful. 

364
00:19:28,700 --> 00:19:30,900
So I've been into some of these 
kind of projects. 

365
00:19:31,200 --> 00:19:33,100
I think I agree with you about 
that problem. 

366
00:19:33,500 --> 00:19:35,700
So you invented this data mesh, 
right? 

367
00:19:35,700 --> 00:19:38,600
So if I may describe data 
mention from your book, you 

368
00:19:38,600 --> 00:19:41,900
mentioned data, mess is a 
decentralized social technical 

369
00:19:41,900 --> 00:19:45,600
approach to share access and 
manage analytical data in 

370
00:19:45,600 --> 00:19:47,700
complex. 
And large-scale environments 

371
00:19:47,700 --> 00:19:50,900
within or across organizations. 
So they have so many interesting

372
00:19:50,900 --> 00:19:53,400
topics but the first that I 
picked is actually you mentioned

373
00:19:53,400 --> 00:19:56,400
it as a decentralized 
socio-technical approach. 

374
00:19:56,700 --> 00:20:00,700
So tell us more about this sure 
it actually started as an 

375
00:20:00,700 --> 00:20:03,900
architecture because I am a 
technologist so I kind of apply 

376
00:20:03,900 --> 00:20:06,000
the lens of technology to solve 
problems. 

377
00:20:06,000 --> 00:20:10,400
So I saw this as really an 
architecture to organize how we 

378
00:20:10,400 --> 00:20:14,800
decouple and how we break down 
these big problem of how I get 

379
00:20:14,800 --> 00:20:18,600
value from data. 
But very quickly, I realized as 

380
00:20:18,600 --> 00:20:24,000
we know, Conway's law and just 
real life experience, technology

381
00:20:24,000 --> 00:20:28,700
and architecture mirrors and get
influenced by the way, we 

382
00:20:28,700 --> 00:20:31,300
organize our organizations and 
teams. 

383
00:20:31,500 --> 00:20:34,000
So, very quickly, I had to like,
self-correct. 

384
00:20:34,000 --> 00:20:36,000
No, this is not just a technical
solution. 

385
00:20:36,000 --> 00:20:37,600
No. 
This is not just an object, or 

386
00:20:37,600 --> 00:20:40,500
we've got to rethink the 
organization of teams. 

387
00:20:40,500 --> 00:20:43,600
The modes of communications, the
contract for data sharing 

388
00:20:43,600 --> 00:20:47,300
between the teams and the She is
like data product. 

389
00:20:47,300 --> 00:20:49,900
Owner was a new role that we 
introduced. 

390
00:20:50,100 --> 00:20:52,000
So parents, it became a 
substitute technical. 

391
00:20:52,000 --> 00:20:57,200
As in we try to find excellence 
in our Solutions involving the 

392
00:20:57,200 --> 00:21:01,900
interaction of people and teams 
and the technology, some people 

393
00:21:01,900 --> 00:21:04,800
say, always a techno social or 
is it Associated that I don't 

394
00:21:04,800 --> 00:21:07,100
really care which one comes 
first, as long as they're both 

395
00:21:07,100 --> 00:21:11,000
involved, hence the word. 
Thanks for sharing that. 

396
00:21:11,000 --> 00:21:12,600
It seems like the Conway's 
losses. 

397
00:21:12,600 --> 00:21:16,000
Really like true principle in 
many software. 

398
00:21:16,200 --> 00:21:18,600
Architecture, right? 
So I think database is probably 

399
00:21:18,600 --> 00:21:22,100
one of it and you mentioned that
it is an approach to solve 

400
00:21:22,100 --> 00:21:24,700
complex and large scale data 
problems. 

401
00:21:25,000 --> 00:21:28,300
So does it mean that not 
everyone will need to go to data

402
00:21:28,300 --> 00:21:31,600
mesh since the beginning? 
Yeah I think at this point in 

403
00:21:31,600 --> 00:21:35,200
time I mean I answered usually 
this question by saying well at 

404
00:21:35,200 --> 00:21:38,200
this point in time, if you don't
have the organizational 

405
00:21:38,200 --> 00:21:41,800
complexity, if you're late for 
laid out so where does model is 

406
00:21:41,800 --> 00:21:45,000
not a bottleneck for you, is the
centralized data thing is doing 

407
00:21:45,000 --> 00:21:47,100
a great job and Everyone's 
happy. 

408
00:21:47,300 --> 00:21:50,600
Well, why introduce a concept? 
That's rather complex. 

409
00:21:51,000 --> 00:21:53,100
It creates kind of system 
complexity. 

410
00:21:53,300 --> 00:21:57,200
So yes, the short answer is. 
Yes it's not for everyone at 

411
00:21:57,200 --> 00:22:01,600
this point to tell her I mean 
future technology advances are 

412
00:22:01,600 --> 00:22:05,000
sinking or approaches are 
process advances in a way that 

413
00:22:05,000 --> 00:22:08,700
bootstrapping with data mesh is 
as easy as bootstrapping or even

414
00:22:08,700 --> 00:22:11,600
easier than its centralized. 
So then at that point you said 

415
00:22:11,600 --> 00:22:13,900
well, there's missions for 
everyone, because the maturity 

416
00:22:13,900 --> 00:22:16,600
of support of the environment 
has reached level of Shorty. 

417
00:22:17,200 --> 00:22:20,100
So you have shared all these 
problems challenges that you saw

418
00:22:20,100 --> 00:22:22,100
before, and you came up with 
this concept. 

419
00:22:22,100 --> 00:22:24,900
But for those people who are 
already in this state of 

420
00:22:24,900 --> 00:22:27,900
complexity of dealing with their
data, either, the data 

421
00:22:27,900 --> 00:22:30,100
architecture pipelines, and 
things like that on 

422
00:22:30,100 --> 00:22:33,300
organization, that is very large
scale, maybe Global where they 

423
00:22:33,300 --> 00:22:36,900
have all this data challenges. 
If you tell them, what are some 

424
00:22:36,900 --> 00:22:40,100
of the reasons why they should 
consider moving to data mesh? 

425
00:22:40,300 --> 00:22:43,400
So maybe in business value or 
maybe in some kind of more 

426
00:22:43,400 --> 00:22:45,600
impact, value-driven, kind of a 
benefits. 

427
00:22:46,300 --> 00:22:50,300
Yeah, just simply if they wanted
subtitle of my book, if they 

428
00:22:50,300 --> 00:22:54,700
want to get value from their 
data, generate data driven value

429
00:22:55,000 --> 00:22:59,900
and they want to do that while I
am applying analytics and AI in 

430
00:23:00,000 --> 00:23:02,700
almost every aspect of their 
business. 

431
00:23:03,200 --> 00:23:07,100
And they want to you to do that.
They need to utilize data for 

432
00:23:07,100 --> 00:23:10,200
all aspects, and all touch 
points and all applications 

433
00:23:10,500 --> 00:23:15,000
inside the company and outside, 
if they have such a mission and 

434
00:23:15,100 --> 00:23:18,600
they want to do that, Reliably 
resilient lie and do that at 

435
00:23:18,600 --> 00:23:21,700
scale fast. 
Then they've got to consider 

436
00:23:21,700 --> 00:23:25,300
this fish, it's all about really
are locking the value of the 

437
00:23:25,300 --> 00:23:27,800
data. 
So, let's give a real world 

438
00:23:27,800 --> 00:23:30,600
example. 
If you are in a particular part 

439
00:23:30,600 --> 00:23:34,000
of the business, let's say I use
this example in my book of this,

440
00:23:34,000 --> 00:23:37,300
but if I like company recorded, 
a think, it's a Digital 

441
00:23:37,300 --> 00:23:39,700
streaming company. 
And if you have a team, whose 

442
00:23:39,700 --> 00:23:44,400
job is really to create 
immersive musical experiences 

443
00:23:44,400 --> 00:23:48,000
personalized for every moments 
of every person in the world 

444
00:23:48,000 --> 00:23:51,500
depending of what they do. 
That team constantly comes up 

445
00:23:51,500 --> 00:23:55,600
with new hypothesis and how to 
use data about music and artists

446
00:23:55,600 --> 00:23:59,000
and listeners and their behavior
to create a more immersive 

447
00:23:59,000 --> 00:24:01,000
experience. 
More personalized that moments 

448
00:24:01,000 --> 00:24:04,400
in life, every one of those 
hypotheses they require 

449
00:24:04,400 --> 00:24:07,000
discovery of the data and access
to the data. 

450
00:24:07,000 --> 00:24:10,900
So are they going to be more 
successful to be able to 

451
00:24:10,900 --> 00:24:14,300
discover and get access to the 
data and even ask people to 

452
00:24:14,300 --> 00:24:15,900
provide the data to data is not 
there. 

453
00:24:16,500 --> 00:24:19,500
If they were working in a 
peer-to-peer fashion, or are 

454
00:24:19,500 --> 00:24:20,900
they going to be more 
successful? 

455
00:24:20,900 --> 00:24:23,100
If it was a centralized team in 
between all parts of the 

456
00:24:23,108 --> 00:24:25,100
business? 
So as an example, if the 

457
00:24:25,100 --> 00:24:29,100
playlist team that generous 
Emerson playlist, want to create

458
00:24:29,100 --> 00:24:32,400
targeted music for people, if 
there are doing cycling or 

459
00:24:32,400 --> 00:24:35,700
running, are they going to be 
more successful to go and talk 

460
00:24:35,700 --> 00:24:39,000
to teams that are taking care of
partnership with cycling 

461
00:24:39,000 --> 00:24:42,300
platforms directly and say, 
look, we need to see what people

462
00:24:42,300 --> 00:24:44,300
are responding to it when 
they're on that pallet on this. 

463
00:24:44,300 --> 00:24:47,600
I suppose as an example, Oil or 
are they successful? 

464
00:24:47,600 --> 00:24:51,600
If they say to a middleman and 
data broker team and say look I 

465
00:24:51,600 --> 00:24:54,700
have this hypothesis. 
So as all of these other teams 

466
00:24:54,900 --> 00:24:58,800
that you need to Now put on your
centralized backlog and plan 

467
00:24:58,800 --> 00:25:03,200
somewhere and get me to the data
that I need, that doesn't scale.

468
00:25:03,400 --> 00:25:07,200
So imagine your organization 
emerging the missions and the 

469
00:25:07,200 --> 00:25:10,300
values that can be enabled 
through the data and see if you 

470
00:25:10,300 --> 00:25:12,600
have bottlenecks that need to be
addressed. 

471
00:25:12,800 --> 00:25:16,200
Then if you do then think about 
dating message though, You 

472
00:25:16,200 --> 00:25:19,100
describe this use case is very 
interesting because yeah, maybe 

473
00:25:19,100 --> 00:25:22,400
not all organizations are in 
this state where you have data 

474
00:25:22,400 --> 00:25:26,500
and you do Discovery and maybe 
shape the next set of data that 

475
00:25:26,500 --> 00:25:30,400
the application produced, again,
do hypothesis and maybe analyze.

476
00:25:30,600 --> 00:25:33,900
And then again, reshape the data
over and over iteratively. 

477
00:25:34,200 --> 00:25:36,800
And then maybe one day you will 
come up with a new insights and 

478
00:25:36,800 --> 00:25:38,900
maybe new business lines as 
well. 

479
00:25:39,100 --> 00:25:42,900
Because the data has transformed
so much with the scale of the 

480
00:25:42,900 --> 00:25:46,100
discovery, and also the scale of
the hypothesis, that the The 

481
00:25:46,100 --> 00:25:48,500
team does. 
So I think that's really a very 

482
00:25:48,500 --> 00:25:51,000
interesting concept. 
I haven't experienced it myself 

483
00:25:51,000 --> 00:25:53,700
because I haven't work in this 
kind of organization, but thanks

484
00:25:53,700 --> 00:25:56,200
for sharing that. 
So let's move on to the 

485
00:25:56,200 --> 00:25:58,500
principles. 
I think, when I read all these 

486
00:25:58,500 --> 00:26:01,500
data mesh concept that you 
share, you always come up with 

487
00:26:01,500 --> 00:26:03,600
this four principles of data 
mesh. 

488
00:26:03,900 --> 00:26:06,600
And when I read all of them, I 
find it interesting because you 

489
00:26:06,600 --> 00:26:10,700
kind of like use other kind of 
framework or maybe approach from

490
00:26:10,700 --> 00:26:14,400
application development and mix 
it into the approach, dealing 

491
00:26:14,400 --> 00:26:16,700
with data. 
So maybe if we can Go briefly 

492
00:26:16,700 --> 00:26:18,800
one by one, right? 
The first one is you call it. 

493
00:26:18,800 --> 00:26:21,500
The principle of two main 
ownership and this is something 

494
00:26:21,500 --> 00:26:24,000
like applying domain-driven 
design to data. 

495
00:26:24,300 --> 00:26:26,700
Maybe if you can tell us more 
about this principle. 

496
00:26:27,500 --> 00:26:30,300
Sure you arrive. 
Your observation is very 

497
00:26:30,300 --> 00:26:32,800
correct. 
That I wasn't as clever as 

498
00:26:32,800 --> 00:26:35,700
creative. 
I basically contextualize the 

499
00:26:35,700 --> 00:26:40,200
things that I had seen working 
in 24 years of my career in 

500
00:26:40,200 --> 00:26:43,100
complex environments in 
operational systems and say, 

501
00:26:43,300 --> 00:26:45,800
let's contextualize and let's 
apply them to the world of 

502
00:26:45,800 --> 00:26:47,400
dating. 
As we've seen these principles 

503
00:26:47,400 --> 00:26:51,300
work before, why shouldn't they 
work with this bottleneck? 

504
00:26:51,400 --> 00:26:52,700
If sole purpose? 
Not only? 

505
00:26:52,900 --> 00:26:56,500
So, the wind ownership is about,
it's the same as what domain 

506
00:26:56,500 --> 00:26:59,300
driven design. 
Meant at the Strategic design 

507
00:26:59,300 --> 00:27:04,100
level applications, which is you
have this smaller business 

508
00:27:04,100 --> 00:27:08,500
domain oriented teams and groups
of people that are 

509
00:27:08,500 --> 00:27:12,300
collaboratively across, 
functionally working to solve 

510
00:27:12,300 --> 00:27:14,400
business problems with 
technology. 

511
00:27:14,800 --> 00:27:19,100
So there Ultimately responsible 
to develop applications and 

512
00:27:19,100 --> 00:27:21,600
software to enable that business
outcome, but they're also 

513
00:27:21,600 --> 00:27:25,700
responsible for using and 
sharing data out for analytical 

514
00:27:25,700 --> 00:27:28,100
purposes. 
For application of machine, 

515
00:27:28,100 --> 00:27:31,400
learning model, again, back to 
the Show streaming, if you have 

516
00:27:31,400 --> 00:27:35,700
a team that is working on your 
player application and their job

517
00:27:35,700 --> 00:27:39,200
is to give the best digital 
experience to a user. 

518
00:27:39,200 --> 00:27:43,600
That's playing music or playing 
and liking it recording whatever

519
00:27:43,600 --> 00:27:45,900
the interactions are. 
You are also responsible 

520
00:27:46,000 --> 00:27:49,300
Responsible as a teen. 
Well, augmented for house, but 

521
00:27:49,300 --> 00:27:53,800
new roles and team members for 
sharing that data in a way that 

522
00:27:53,800 --> 00:27:58,500
data can be used to directly by 
some sort of an analytical work 

523
00:27:58,500 --> 00:28:01,600
load, and that analytical work 
of might be a machine learning 

524
00:28:01,600 --> 00:28:06,100
model, that is being trained by 
your data and some other data 

525
00:28:06,100 --> 00:28:09,400
from other domains or it could 
be the reports that we are 

526
00:28:09,400 --> 00:28:13,800
producing in terms of errors and
anomalies of this application. 

527
00:28:13,800 --> 00:28:17,400
So we can improve it over time. 
That's the core of it is that 

528
00:28:17,400 --> 00:28:21,700
break down the responsibility of
data, sharing around the seams 

529
00:28:21,700 --> 00:28:24,900
of organization. 
So you have this infinitely, 

530
00:28:24,900 --> 00:28:27,500
scalable model. 
As you introduce new domains, 

531
00:28:27,500 --> 00:28:30,800
you introduce new data sets that
there was no mates can use and 

532
00:28:30,800 --> 00:28:33,500
share and give the 
responsibility of this charity 

533
00:28:33,500 --> 00:28:36,000
to people for analytics. 
Guessing it for this, kind of 

534
00:28:36,000 --> 00:28:38,900
cross-cutting use cases to 
people that are capable to be 

535
00:28:38,900 --> 00:28:41,400
responsible for it because 
they're so close to it. 

536
00:28:41,400 --> 00:28:44,100
They understand, they know what 
they say about. 

537
00:28:44,200 --> 00:28:45,900
Don't give that responsibility 
to someone. 

538
00:28:46,000 --> 00:28:48,700
One Downstream that actually 
doesn't know the domain and 

539
00:28:48,700 --> 00:28:51,600
spirit hard for them to keep 
cognitive load of knowing all 

540
00:28:51,600 --> 00:28:53,400
the domains is their team's 
heads. 

541
00:28:53,600 --> 00:28:56,600
So that's the first principle. 
The way you describe it? 

542
00:28:56,600 --> 00:28:58,500
It sounds intuitive, right? 
Yeah. 

543
00:28:58,500 --> 00:29:00,800
Why not? 
But we came from the traditional

544
00:29:00,800 --> 00:29:03,100
approach where we have 
centralized team, they have to 

545
00:29:03,100 --> 00:29:06,600
understand all the domain within
the company organization, 

546
00:29:06,700 --> 00:29:10,800
understand the data model, the 
evolution of all the data, and 

547
00:29:10,800 --> 00:29:14,100
try to put them in one central 
place, but I think your approach

548
00:29:14,100 --> 00:29:17,400
here they are principal putting 
in To a domain ownership means 

549
00:29:17,400 --> 00:29:21,100
that the domain team themselves 
is responsible for, not just the

550
00:29:21,100 --> 00:29:24,300
operational data, but also the 
analogy call data part where 

551
00:29:24,300 --> 00:29:27,800
probably they will transform the
operational data and responsible

552
00:29:27,800 --> 00:29:29,900
for sharing them for analytical 
purposes. 

553
00:29:29,900 --> 00:29:33,700
Well, and since they are the 
experts of the domain, they are 

554
00:29:33,700 --> 00:29:36,700
probably the best person to come
up with that kind of a data 

555
00:29:36,700 --> 00:29:39,300
model. 
So I think it's really intuitive

556
00:29:39,300 --> 00:29:42,800
after you explain it. 
There's also one concept that is

557
00:29:42,800 --> 00:29:44,800
on this principle, right? 
Where you mentioned. 

558
00:29:44,800 --> 00:29:48,700
That the data pipeline, Is now 
is not a responsibility of a 

559
00:29:48,708 --> 00:29:51,400
central theme, but now it 
becomes an internal 

560
00:29:51,400 --> 00:29:54,100
implementation of that domain 
team itself. 

561
00:29:54,300 --> 00:29:57,200
So maybe if you can describe why
this is so important. 

562
00:29:57,500 --> 00:30:01,100
Sure, well, if there's a fish is
really successful. 

563
00:30:01,100 --> 00:30:04,700
The way I had imagined that 
they're no longer be any data 

564
00:30:04,700 --> 00:30:07,900
pipelines. 
I think if you go to the macro 

565
00:30:07,900 --> 00:30:10,500
level, macro view of your 
architecture, you really 

566
00:30:10,500 --> 00:30:12,700
shouldn't see data pipelines 
anymore. 

567
00:30:12,700 --> 00:30:15,800
Between these domain between the
data products. 

568
00:30:16,000 --> 00:30:18,400
Just concept that we can 
introduce it, the next 

569
00:30:18,400 --> 00:30:21,300
principle. 
So if you think about pie, that 

570
00:30:21,300 --> 00:30:24,700
is what's the purpose of them 
pipelines or job oriented 

571
00:30:24,700 --> 00:30:29,200
task-oriented, computations that
happen on some input from the 

572
00:30:29,200 --> 00:30:33,100
data and transforms and put it 
on some output sink and you 

573
00:30:33,100 --> 00:30:36,800
repeat these task oriented 
process until the leader gets 

574
00:30:36,800 --> 00:30:38,800
transferred. 
In a mode that somebody can use 

575
00:30:38,800 --> 00:30:40,800
it. 
Usually there is structure 

576
00:30:40,800 --> 00:30:43,300
around looking. 
We need to extract information. 

577
00:30:43,300 --> 00:30:45,600
We need to cleanse them with a 
to model. 

578
00:30:45,900 --> 00:30:49,600
And the usually done in between 
the scenes of the data outside 

579
00:30:49,600 --> 00:30:52,600
of the source outside of the 
destination somewhere in 

580
00:30:52,600 --> 00:30:54,600
between. 
So, Taylor is completely 

581
00:30:54,600 --> 00:30:57,900
challenges that concept, because
we're no longer working this 

582
00:30:57,900 --> 00:30:59,900
task oriented kind of 
environment. 

583
00:30:59,900 --> 00:31:02,800
We working on this value 
oriented outcome oriented 

584
00:31:02,800 --> 00:31:05,200
environment data product re into
the environment. 

585
00:31:05,400 --> 00:31:08,800
So then the job of the 
transformation is really an 

586
00:31:08,800 --> 00:31:12,200
implementation details. 
One of these data products in 

587
00:31:12,200 --> 00:31:15,800
one domain, it's not something 
that happens in between. 

588
00:31:16,300 --> 00:31:20,800
And it's really falls into the 
similar principle of micro 

589
00:31:20,800 --> 00:31:23,800
services that we had in 
microservices world. 

590
00:31:23,800 --> 00:31:26,900
This Enterprise service was an 
anti-pattern, right? 

591
00:31:26,900 --> 00:31:30,700
Because we want to localize 
logic and computation and 

592
00:31:30,700 --> 00:31:35,500
complexity inside and boundary 
of a contract abstracted within 

593
00:31:35,500 --> 00:31:38,000
a service, the Enterprise 
service bus wasn't doing that. 

594
00:31:38,000 --> 00:31:41,700
So, we came up with this idea of
smart, endpoints and dumb pipes.

595
00:31:42,000 --> 00:31:45,700
So your pipes are super down 
there, just transform data and 

596
00:31:45,700 --> 00:31:48,800
your Our inputs are smart 
because they implement the logic

597
00:31:48,800 --> 00:31:51,600
behind the apis of that. 
Led to kind of more API thinking

598
00:31:51,600 --> 00:31:53,300
world. 
It's the same process. 

599
00:31:53,400 --> 00:31:57,400
So these pipelines are like the 
Enterprise service bus analogy. 

600
00:31:57,400 --> 00:31:59,700
If I choose and they have the 
same challenges, so we've got to

601
00:31:59,708 --> 00:32:02,000
break them up. 
Followed the pieces that are 

602
00:32:02,000 --> 00:32:05,000
relevant, should feel where they
should be known, they can be 

603
00:32:05,000 --> 00:32:07,100
implemented as Parkland if you 
want to within the 

604
00:32:07,100 --> 00:32:10,900
implementation or other ways. 
Again, very intuitive. 

605
00:32:10,900 --> 00:32:13,700
If you again compared with the 
application development, right? 

606
00:32:13,700 --> 00:32:16,900
So yes, be has long. 
Been an anti-pattern So the same

607
00:32:16,900 --> 00:32:19,700
concept in data, if you want to 
connect to different data 

608
00:32:19,700 --> 00:32:23,000
products, you should not create 
like a complex data pipelines. 

609
00:32:23,000 --> 00:32:25,400
It should be just maybe 
transferring the data when 

610
00:32:25,400 --> 00:32:27,400
transferring the data. 
This is also another concept 

611
00:32:27,400 --> 00:32:31,000
under this principle that you 
mentioned, there is now probably

612
00:32:31,000 --> 00:32:35,000
not a kind of like notion of one
source of truth anymore about 

613
00:32:35,000 --> 00:32:37,800
the data. 
You will have multi maybe shape 

614
00:32:37,800 --> 00:32:40,900
of the data multiple copies. 
Maybe it is in the domain team 

615
00:32:40,900 --> 00:32:43,400
themselves, or maybe it's 
already copied to the other 

616
00:32:43,400 --> 00:32:46,200
consumer of the data where they 
will use it for their the use 

617
00:32:46,200 --> 00:32:48,600
case. 
So why is this the case? 

618
00:32:48,600 --> 00:32:50,500
Why normal single source of 
truth? 

619
00:32:50,700 --> 00:32:53,100
I thought, like in the data 
world, people love to have a 

620
00:32:53,100 --> 00:32:54,700
rest this single source of 
Truth. 

621
00:32:54,900 --> 00:32:58,900
Where is the data that I can 
trust because single source of 

622
00:32:58,900 --> 00:33:01,500
truth? 
It's not the real thing, it just

623
00:33:01,500 --> 00:33:04,400
doesn't exist in real world. 
Okay? 

624
00:33:04,400 --> 00:33:08,700
So let's unpack that. 
I don't intend to claim that we 

625
00:33:08,700 --> 00:33:12,200
are becoming irresponsible about
data and you will have 

626
00:33:12,200 --> 00:33:16,300
contradictory copies of the data
lying around and have Kind of 

627
00:33:16,300 --> 00:33:18,600
concerned those that's not the 
aim. 

628
00:33:18,600 --> 00:33:24,200
The aim is we still want to be 
able to get a consistent View 

629
00:33:24,200 --> 00:33:28,100
and understanding of the data, 
but we want to do that in a way 

630
00:33:28,100 --> 00:33:32,000
that it doesn't slow movement. 
It doesn't snow value 

631
00:33:32,000 --> 00:33:34,200
generation. 
It doesn't become a sterile 

632
00:33:34,200 --> 00:33:38,300
source of Truth, very quickly. 
We want to do that in a way that

633
00:33:38,300 --> 00:33:42,100
we support the chaos of reality 
of organizations, their 

634
00:33:42,100 --> 00:33:45,000
different teams will be 
different spaced, they generate 

635
00:33:45,000 --> 00:33:48,200
different bits, And pieces of 
properties of the same entity, 

636
00:33:48,200 --> 00:33:50,500
but those properties come from 
different sources with different

637
00:33:50,500 --> 00:33:52,700
Cadence. 
So we want to embrace that 

638
00:33:52,700 --> 00:33:57,600
almost complexity chaos, but yet
create a system that gives the 

639
00:33:57,600 --> 00:34:00,400
same outcome that the single 
source of Truth wants to give, 

640
00:34:00,500 --> 00:34:05,700
which is if I search for 
information about customer even 

641
00:34:05,700 --> 00:34:08,900
though that information comes 
from different places, I can 

642
00:34:08,900 --> 00:34:12,900
understand a particular snapshot
of the customer at a point in 

643
00:34:12,900 --> 00:34:15,699
time. 
Has a set of consistent values. 

644
00:34:16,100 --> 00:34:19,500
That's why I challenge this 
notion of single source of 

645
00:34:19,500 --> 00:34:22,100
Truth. 
So they admit when people don't 

646
00:34:22,100 --> 00:34:25,600
actually read or understand, it 
has a lot of constraints and 

647
00:34:25,600 --> 00:34:29,900
disciplines built into it. 
For example, the data that data 

648
00:34:29,900 --> 00:34:34,199
nodes provide or read-only, they
never change their temporal, 

649
00:34:34,199 --> 00:34:36,500
they have two x terms by 
temporal. 

650
00:34:36,699 --> 00:34:40,000
So at any point in time, there 
are streamed across different 

651
00:34:40,000 --> 00:34:43,500
nodes in a way that is some data
arrives, new data arrives 

652
00:34:43,500 --> 00:34:47,000
Upstream, the downstream notes 
that are For transforming that 

653
00:34:47,000 --> 00:34:49,600
were copying that and 
transforming it into a new shape

654
00:34:49,600 --> 00:34:54,100
of the data they get notified. 
They have a responsibility to 

655
00:34:54,100 --> 00:34:57,800
either react on it and generate 
a new slice of the database 

656
00:34:57,800 --> 00:35:00,900
pointy toilet or not. 
And then also this image 

657
00:35:00,900 --> 00:35:04,700
provides a means of stitching is
probably same Concepts like a 

658
00:35:04,707 --> 00:35:07,400
customer that comes from the 
call center versus customer. 

659
00:35:07,400 --> 00:35:10,600
That comes from the cover set, 
for the punishing concept of the

660
00:35:10,600 --> 00:35:14,500
customer, can be stitched 
together by the consumer because

661
00:35:14,500 --> 00:35:17,200
there are links that are It's 
between those systems. 

662
00:35:17,200 --> 00:35:20,800
So between those little product.
So there are set of constraints.

663
00:35:21,100 --> 00:35:24,100
There are set of operational 
disciplines like this policy, 

664
00:35:24,100 --> 00:35:28,500
meaning cage, and by temporality
and immutable data, but still 

665
00:35:28,500 --> 00:35:31,900
results in the saying, outcome 
of a single source of truth. 

666
00:35:32,000 --> 00:35:35,900
But it's designed for inherently
complex, business model, and 

667
00:35:35,900 --> 00:35:37,700
operating model, if that makes 
sense. 

668
00:35:38,400 --> 00:35:40,600
And you mentioned it as a most 
relevant copyright. 

669
00:35:40,600 --> 00:35:42,900
So maybe you don't get the 
latest up-to-date. 

670
00:35:42,900 --> 00:35:46,500
So it's like the concept of this
asynchronous or maybe eventual 

671
00:35:46,500 --> 00:35:49,300
consistency. 
Maybe you just need a snapshot 

672
00:35:49,300 --> 00:35:50,900
of a data at certain point in 
time. 

673
00:35:51,200 --> 00:35:52,900
And yet the consumer will 
decide, okay? 

674
00:35:52,900 --> 00:35:55,900
I just need this kind of data 
instead of always getting the 

675
00:35:55,900 --> 00:35:58,600
latest. 
Let's move on to the next 

676
00:35:58,600 --> 00:36:01,600
principle which is a principle 
of data as a product. 

677
00:36:01,600 --> 00:36:05,000
And I see that you are applying 
product thinking to this 

678
00:36:05,000 --> 00:36:07,300
principle. 
So tell us more why. 

679
00:36:07,300 --> 00:36:09,900
It is important to treat data is
a product. 

680
00:36:10,500 --> 00:36:14,300
Yeah, I think of Mind shift that
needs to happen when we put data

681
00:36:14,300 --> 00:36:18,200
sharing and serving it, Lighting
the experience of people using 

682
00:36:18,200 --> 00:36:22,900
data as a first-class concern, 
is also an antidote to the first

683
00:36:22,900 --> 00:36:26,300
verse of all the problems are 
the first principle, so domain 

684
00:36:26,500 --> 00:36:31,200
oriented data ownership. 
One can imagine can lead to Beta

685
00:36:31,200 --> 00:36:34,200
siloing on the player domain, 
but the data that I need to 

686
00:36:34,200 --> 00:36:36,900
improve my application. 
So why should I care about 

687
00:36:36,900 --> 00:36:40,200
sharing that and be responsible 
for consumers, which adds a ton 

688
00:36:40,200 --> 00:36:42,300
of work. 
So this has a product try to 

689
00:36:42,300 --> 00:36:45,700
incentivize people to share that
data. 

690
00:36:45,800 --> 00:36:49,600
Be part of an ecosystem that is 
generating value through the 

691
00:36:49,600 --> 00:36:51,600
exchange and through data 
sharing. 

692
00:36:51,800 --> 00:36:56,300
And again, put some discipline 
and constraints in place for 

693
00:36:56,300 --> 00:37:00,800
that to be done effectively. 
So defines such roles of people 

694
00:37:00,800 --> 00:37:04,300
that are responsible, for that 
define success, metrics for data

695
00:37:04,300 --> 00:37:06,200
as a product. 
Define, an architectural 

696
00:37:06,200 --> 00:37:09,600
consent, if we define usability 
characteristics around 

697
00:37:09,600 --> 00:37:12,000
discoverability address, but 
it's like all the things that 

698
00:37:12,000 --> 00:37:14,800
make the experience of the data 
user, really easy, really 

699
00:37:14,800 --> 00:37:17,700
delightful. 
Those need to be translated into

700
00:37:17,700 --> 00:37:20,800
structural Architectural 
Components that are built into 

701
00:37:20,800 --> 00:37:23,600
this data as a product. 
So then you have kind of 

702
00:37:23,600 --> 00:37:26,100
technology that needs to shift 
and change. 

703
00:37:26,300 --> 00:37:28,000
So, yeah. 
So in short is an antidote to 

704
00:37:28,000 --> 00:37:30,200
the problems that are always 
from the first principle and 

705
00:37:30,200 --> 00:37:35,600
also really focus on again 
getting value from data who gets

706
00:37:35,600 --> 00:37:39,000
value from data that users do. 
So, let's put the first. 

707
00:37:39,700 --> 00:37:43,700
I like the way you explain, why 
the concept of a product is up 

708
00:37:43,700 --> 00:37:46,900
here because you mentioned for a
successful Products, you need 

709
00:37:46,900 --> 00:37:50,800
these three attributes which is 
visibility valuable and usable 

710
00:37:51,100 --> 00:37:53,400
and I think traditionally again,
we just read data. 

711
00:37:53,400 --> 00:37:55,400
Okay. 
This is just a data you go and 

712
00:37:55,400 --> 00:37:57,900
figure it out so sometimes it's 
not usable. 

713
00:37:57,900 --> 00:38:00,300
So sometimes maybe the query 
language is different. 

714
00:38:00,300 --> 00:38:02,600
The database Technologies 
different because we all have 

715
00:38:02,600 --> 00:38:06,200
this polyglot database 
Technologies or maybe it's so 

716
00:38:06,200 --> 00:38:08,200
ancient right Legacy 
Technologies. 

717
00:38:08,200 --> 00:38:09,700
We don't know how to deal with 
it. 

718
00:38:09,900 --> 00:38:12,600
So I think if treating it as a 
product we also need to think 

719
00:38:12,600 --> 00:38:15,700
about the usability aspect. 
So I think that is definitely 

720
00:38:15,800 --> 00:38:19,600
The key when I read about this 
principle, there are also a new 

721
00:38:19,600 --> 00:38:22,800
role that is being created 
because of this concept of data 

722
00:38:22,800 --> 00:38:25,400
is a product. 
You mentioned about data product

723
00:38:25,400 --> 00:38:27,500
owner, and data product 
developer. 

724
00:38:27,700 --> 00:38:30,100
So not all teams have these 
roles yet. 

725
00:38:30,300 --> 00:38:32,800
Tell us more about the 
importance of these two roles. 

726
00:38:33,600 --> 00:38:37,500
Yeah, so if they support it 
becomes a thing that we are 

727
00:38:37,600 --> 00:38:41,800
creating maintaining operating 
evolving retiring with it's 

728
00:38:41,800 --> 00:38:45,600
useless and nobody use it, then 
they have to be people and roll.

729
00:38:45,800 --> 00:38:49,700
Rose to take that responsibility
on and it's very unfair. 

730
00:38:50,000 --> 00:38:54,500
I think it's impossible to say 
to app developers who's working 

731
00:38:54,500 --> 00:38:57,600
consumer or very different 
personas, like they're serving 

732
00:38:57,600 --> 00:39:01,600
the end-user to say, oh now from
here, all of you also have this 

733
00:39:01,600 --> 00:39:04,400
other responsibilities. 
So not only your serving these 

734
00:39:04,400 --> 00:39:07,700
end users that are interacting 
with your player application 

735
00:39:07,700 --> 00:39:10,400
pressing buttons and make sure 
they have a responsive app and 

736
00:39:10,400 --> 00:39:14,600
all of those great things but 
also half of your day you have 

737
00:39:14,600 --> 00:39:17,400
to face around. 
And face this data analyst and 

738
00:39:17,400 --> 00:39:21,100
designs in different domains who
want to use the analytical data 

739
00:39:21,100 --> 00:39:23,700
that you're generating, you've 
got to serve them to it's 

740
00:39:23,700 --> 00:39:25,900
impossible. 
You turned up two bosses like 

741
00:39:25,900 --> 00:39:29,400
because service to purposes at 
the same time so if there are 

742
00:39:29,400 --> 00:39:31,600
super humans that can do both 
jobs. 

743
00:39:31,600 --> 00:39:32,900
So be it. 
That's fine. 

744
00:39:32,900 --> 00:39:36,200
Maybe it is possible to kind of 
share your time and split your 

745
00:39:36,200 --> 00:39:38,300
time that way. 
But nevertheless you need to 

746
00:39:38,300 --> 00:39:40,500
have a very explicit 
responsibility and 

747
00:39:40,500 --> 00:39:42,500
accountability for that. 
Part of the job. 

748
00:39:42,700 --> 00:39:44,800
If there are people that are 
appearing, like the data 

749
00:39:44,800 --> 00:39:47,500
product, it will offer and 
adaptable, preparing and of 

750
00:39:47,508 --> 00:39:50,500
collaborating closely for the 
two different people playing 

751
00:39:50,500 --> 00:39:54,300
these roles and that's fine too.
Yeah, so unless these kind of 

752
00:39:54,300 --> 00:39:57,300
idea of the data Pride doesn't 
happen out of a good intention, 

753
00:39:57,300 --> 00:40:00,500
we have to allocate space, we 
have to empower people have to 

754
00:40:00,500 --> 00:40:03,300
keep people accountable as 
follows for hence, the roads. 

755
00:40:04,200 --> 00:40:05,700
Yeah. 
And also not to mention the 

756
00:40:05,700 --> 00:40:07,400
skill set. 
They are totally different 

757
00:40:07,400 --> 00:40:09,200
Technologies, different 
paradigms. 

758
00:40:09,400 --> 00:40:12,200
I think it's very difficult to 
find people who can Master both 

759
00:40:12,200 --> 00:40:14,900
application development and also
data engineering. 

760
00:40:15,200 --> 00:40:18,200
I think for That to be also 
usable as a product that you 

761
00:40:18,200 --> 00:40:20,800
mention about all other 
usability attributes, like 

762
00:40:20,800 --> 00:40:23,200
discoverability, address 
ability. 

763
00:40:23,200 --> 00:40:26,600
It should also be understandable
and it's trustworthy as well, 

764
00:40:26,600 --> 00:40:29,900
and it can be interoperable. 
So imagine if you have multiple 

765
00:40:29,900 --> 00:40:32,900
consumers, then they want to 
access your data, but they do 

766
00:40:32,900 --> 00:40:35,900
have set of constraints on how 
they would integrate with your 

767
00:40:35,900 --> 00:40:38,100
data. 
I think interoperability is also

768
00:40:38,100 --> 00:40:41,700
a good concern that the data 
domain owner should think about.

769
00:40:41,700 --> 00:40:44,600
And that's why I think having a 
data product owner that maybe 

770
00:40:44,600 --> 00:40:47,500
can Define the set of Of 
requirements and maybe it's kind

771
00:40:47,500 --> 00:40:51,200
of all usability concerns. 
I think that is key why these 

772
00:40:51,200 --> 00:40:53,700
rules exist. 
So let's move on maybe to the 

773
00:40:53,700 --> 00:40:56,900
next principle which is about 
principle of the self-serve data

774
00:40:56,900 --> 00:40:59,400
platform. 
I think this is also interesting

775
00:40:59,400 --> 00:41:03,300
because you kind of like apply 
platform thinking to data mesh, 

776
00:41:03,700 --> 00:41:06,200
why we should have a self-serve 
data platform. 

777
00:41:06,900 --> 00:41:10,800
Yeah, I think it's an obvious 
one, but let's go deeper into it

778
00:41:10,800 --> 00:41:14,400
and say, maybe answer it. 
What is this little platform? 

779
00:41:14,600 --> 00:41:17,900
So, when we think about It's the
roles of Platforms in general 

780
00:41:17,900 --> 00:41:20,600
platforms. 
Are often shirt kind of 

781
00:41:20,600 --> 00:41:24,300
infrastructure on top of which 
you build domain specific 

782
00:41:24,300 --> 00:41:25,800
solution. 
So they often like the Mets 

783
00:41:25,800 --> 00:41:30,300
diagnostic infrastructure that 
Empower other teams to build 

784
00:41:30,400 --> 00:41:33,200
Solutions on top. 
So the valley horizontal not 

785
00:41:33,200 --> 00:41:36,700
really bear to go so much. 
So in an organization, that's 

786
00:41:36,700 --> 00:41:39,600
implementing data mesh. 
There you have this domain 

787
00:41:39,600 --> 00:41:41,600
teams, they have active faults 
in them. 

788
00:41:41,600 --> 00:41:44,000
You have data product, folks in 
them and they're doing their 

789
00:41:44,000 --> 00:41:48,100
daily job there. 
Daily job should be focused on 

790
00:41:48,100 --> 00:41:51,100
delivering value based on the 
outcomes of that domain. 

791
00:41:51,400 --> 00:41:55,200
Their daily job should not be 
focused on metal work, creating 

792
00:41:55,200 --> 00:41:58,200
kind of the way diagnostic 
pieces of technology that they 

793
00:41:58,200 --> 00:42:01,500
need. 
So, I think platforms as mean to

794
00:42:01,600 --> 00:42:05,900
empower autonomous teams to 
lower, their cognitive load to 

795
00:42:05,900 --> 00:42:08,200
do what they need to do. 
More easily. 

796
00:42:08,400 --> 00:42:09,700
They're wonderful. 
They're necessary. 

797
00:42:10,000 --> 00:42:13,200
In terms of data mesh, the data 
platforms, many of them exist or

798
00:42:13,200 --> 00:42:18,000
lots of Technology out there. 
They're built To give basic 

799
00:42:18,000 --> 00:42:22,200
tools like you want story short?
I will give you a storage. 

800
00:42:22,200 --> 00:42:25,300
I can provision that bright. 
You want workflow processing. 

801
00:42:25,400 --> 00:42:28,200
That's why like I'll give you 
that but the to eat slow devil 

802
00:42:28,200 --> 00:42:31,700
for a developing the product. 
So we need a new layer of the 

803
00:42:31,700 --> 00:42:35,700
platform that really takes data 
product away at least a to mesh 

804
00:42:35,700 --> 00:42:40,200
that I defined envisions treated
as a first-class concern and 

805
00:42:40,200 --> 00:42:43,000
hides away. 
Those details of, oh, I need a 

806
00:42:43,000 --> 00:42:47,000
storage partnered with a, see 
that and it gives His life to a 

807
00:42:47,000 --> 00:42:50,400
completely new concept, this new
concept of this Quantum as it 

808
00:42:50,400 --> 00:42:52,900
did product. 
So the job of that platform, the

809
00:42:52,900 --> 00:42:57,000
reason I put it in was making it
feasible for independent, domain

810
00:42:57,000 --> 00:43:00,100
teams to do data work. 
And some of the attributes that 

811
00:43:00,100 --> 00:43:02,800
you mention of this data 
platform is that it should be 

812
00:43:02,800 --> 00:43:05,800
autonomous interoperable and 
domain Gnostic. 

813
00:43:05,800 --> 00:43:08,300
So I think one of the challenges
when I was working with data 

814
00:43:08,400 --> 00:43:11,400
related stuff, also is that? 
Yeah you have all these tools 

815
00:43:11,400 --> 00:43:14,300
like you mentioned, but 
bootstrapping all these tools 

816
00:43:14,300 --> 00:43:17,200
takes a long time. 
You have the setup clusters, you

817
00:43:17,200 --> 00:43:20,500
have to maybe install things 
dependencies and on top of that,

818
00:43:20,500 --> 00:43:23,900
then you have to write code, you
have to maybe understand the 

819
00:43:23,900 --> 00:43:25,800
source, the sink and things like
that. 

820
00:43:26,000 --> 00:43:28,500
And then you have to write the 
pipeline's itself and then 

821
00:43:28,500 --> 00:43:31,200
deploy it, and things like that.
Yeah, it takes a lot of effort 

822
00:43:31,500 --> 00:43:33,500
just to come up with a very 
simple Pipeline. 

823
00:43:33,900 --> 00:43:36,500
And I could imagine having this 
kind of self-serve data 

824
00:43:36,500 --> 00:43:39,300
platform, maybe something like 
the UI console, where you can 

825
00:43:39,300 --> 00:43:41,200
just go login. 
And, you know, click, I want 

826
00:43:41,200 --> 00:43:44,900
this data from this source and 
move it to my sink and then it 

827
00:43:44,900 --> 00:43:47,700
just creates everything for you.
I think that will be a perfect 

828
00:43:47,700 --> 00:43:50,900
scenario where maybe some of the
tools, not yet, catching up. 

829
00:43:51,100 --> 00:43:54,600
But yeah, hopefully, one day we 
will reach the experience for 

830
00:43:54,600 --> 00:43:57,500
the data engineer or maybe for 
business people, they don't even

831
00:43:57,500 --> 00:44:00,400
need to care about dealing with 
data Engineers themselves. 

832
00:44:00,800 --> 00:44:03,500
So, you mentioned about this 
concept of self-serve data 

833
00:44:03,500 --> 00:44:06,800
platform because one of the 
challenges of building this 

834
00:44:06,800 --> 00:44:09,900
platform is that you will need 
to make it agnostic. 

835
00:44:09,900 --> 00:44:12,600
So, I think this is probably one
of the challenges because we 

836
00:44:12,600 --> 00:44:14,500
have so many different data 
Technologies. 

837
00:44:14,800 --> 00:44:17,800
So how Should we think about 
building this platform? 

838
00:44:17,800 --> 00:44:20,100
So that it becomes agnostic 
because we have so many 

839
00:44:20,100 --> 00:44:22,800
Technologies, right? 
We have so many different shape 

840
00:44:22,800 --> 00:44:25,700
of database Technologies. 
Yeah, I mean it depends the 

841
00:44:25,700 --> 00:44:29,100
agnostics city, I guess of the 
platform and its independence 

842
00:44:29,100 --> 00:44:32,500
from underlying kind of 
Technologies the level of it 

843
00:44:32,500 --> 00:44:35,300
depends on the appetites. 
It's an organization that how 

844
00:44:35,300 --> 00:44:39,200
independent they want to be. 
So I think what we need is 

845
00:44:39,200 --> 00:44:43,100
interoperability between the 
different technology, so if it's

846
00:44:43,100 --> 00:44:47,900
still the solution on top, And 
the solution requires data from 

847
00:44:47,900 --> 00:44:50,100
across two different platforms. 
There is a level of 

848
00:44:50,100 --> 00:44:53,200
interoperability that I can 
access data across two different

849
00:44:53,200 --> 00:44:55,800
clouds or interest to different 
technology Stacks. 

850
00:44:56,100 --> 00:44:59,300
So at minimum we need and the 
year that creates that 

851
00:44:59,300 --> 00:45:02,000
interoperability even if the 
vendors themselves are not 

852
00:45:02,000 --> 00:45:03,600
incentivized to do that. 
Right. 

853
00:45:03,600 --> 00:45:09,100
Now when it comes to underlying 
infrastructure agnostic again, I

854
00:45:09,107 --> 00:45:12,600
don't think it's meaningful in 
all organizations because you 

855
00:45:12,600 --> 00:45:16,400
end up with a like most common 
denominator of the Features that

856
00:45:16,400 --> 00:45:18,000
are available on all the 
platforms. 

857
00:45:18,000 --> 00:45:21,000
And that's not ideal. 
That's a lot of work and very 

858
00:45:21,000 --> 00:45:24,300
little result so don't know if 
it needs to be completely 

859
00:45:24,300 --> 00:45:27,200
agnostic. 
But we have to have the pieces 

860
00:45:27,200 --> 00:45:30,300
of it, that enables 
interoperability and movement 

861
00:45:30,600 --> 00:45:34,400
moving from one to another and 
remove locking as much as 

862
00:45:34,400 --> 00:45:37,000
possible. 
And those pieces are usually 

863
00:45:37,000 --> 00:45:38,800
around cross-cutting concerns, 
right? 

864
00:45:38,800 --> 00:45:42,900
How do I manage security? 
How do I be able to automation? 

865
00:45:42,900 --> 00:45:45,900
So that if tomorrow I want to 
move to a different Set of 

866
00:45:45,900 --> 00:45:50,000
infrastructure, my processes are
automated and not hand-cranked. 

867
00:45:50,400 --> 00:45:53,000
I can kind of through automation
facilitate the movement, much 

868
00:45:53,000 --> 00:45:56,600
faster, that's the way I think 
about this being Tech agnostic. 

869
00:45:56,600 --> 00:46:00,500
As opposed to a nice layer on 
top, I don't think that's really

870
00:46:00,500 --> 00:46:04,100
realistic speaking about 
cross-cutting concerns. 

871
00:46:04,100 --> 00:46:07,100
So this also touching on the 
next principle. 

872
00:46:07,100 --> 00:46:09,400
The last principle which is 
principle of Federated 

873
00:46:09,400 --> 00:46:12,800
computational governance. 
It's quite a mouthful to mention

874
00:46:12,800 --> 00:46:15,400
that but it's taking care of all
these cross-cutting concerns are

875
00:46:15,700 --> 00:46:19,100
Mention things like for example,
security policies and things 

876
00:46:19,100 --> 00:46:21,300
like that. 
You are kind of like applying 

877
00:46:21,300 --> 00:46:24,300
systems thinking to this 
principle so that we can govern 

878
00:46:24,300 --> 00:46:27,200
the data better. 
So it share more about this 

879
00:46:27,200 --> 00:46:30,700
principle because to some people
this might be hard to kind of 

880
00:46:30,700 --> 00:46:32,600
like understand. 
Yeah. 

881
00:46:32,600 --> 00:46:36,100
It's a mouthful already. 
And if I could sneak in another 

882
00:46:36,100 --> 00:46:39,600
word, I probably have called 
this principle of embittered 

883
00:46:39,600 --> 00:46:44,900
Federated competition covering. 
But I think Marti Fowler would 

884
00:46:44,900 --> 00:46:48,100
have not Not posted my article, 
right, did that? 

885
00:46:48,300 --> 00:46:51,500
Yeah, so I think that the 
concept is really is again, I 

886
00:46:51,500 --> 00:46:54,600
don't to do two problems that 
arise from the previous 

887
00:46:54,600 --> 00:46:57,200
principles, which is we need 
interoperability. 

888
00:46:57,200 --> 00:47:00,700
We have now this despair it sets
of data products domain 

889
00:47:00,700 --> 00:47:03,800
oriented, their own team, their 
own Cadence their life cycle. 

890
00:47:04,100 --> 00:47:09,300
How can we apply set of concerns
that need to be standardized 

891
00:47:09,300 --> 00:47:12,000
across all of them? 
And what's the best way to go 

892
00:47:12,000 --> 00:47:14,500
about defining them? 
What's the best way to go about 

893
00:47:14,500 --> 00:47:17,400
implementing them? 
Observing and enforcing them 

894
00:47:17,600 --> 00:47:21,400
that leads to this principle. 
So as an example, if you know to

895
00:47:21,400 --> 00:47:25,000
have secure data or you need to 
have high quality data, let's go

896
00:47:25,000 --> 00:47:27,300
with high quality. 
So server quality data 

897
00:47:27,300 --> 00:47:30,300
definition of quality and then 
enforcing quality and Sherry 

898
00:47:30,300 --> 00:47:33,500
quality data. 
One way of doing it is say, 

899
00:47:33,800 --> 00:47:35,400
okay. 
I'm going to put a quality 

900
00:47:35,400 --> 00:47:38,000
control team. 
My governance team in the 

901
00:47:38,000 --> 00:47:41,400
process of generating every 
data, and this is going to sit 

902
00:47:41,400 --> 00:47:45,300
in the middle and verify at some
points in that life cycle of the

903
00:47:45,300 --> 00:47:48,000
day. 
Data its beta acceptable to be 

904
00:47:48,000 --> 00:47:50,100
shared. 
That's where system taking comes

905
00:47:50,100 --> 00:47:52,400
to play. 
That system is going to have a 

906
00:47:52,408 --> 00:47:54,900
massive bottlenecks and it's not
going to scale. 

907
00:47:55,000 --> 00:47:59,100
So how can we achieve a defined 
level of quality without 

908
00:47:59,100 --> 00:48:01,500
creating just necessarily just 
controls? 

909
00:48:01,800 --> 00:48:04,900
We need the consensus or 
definition around what 

910
00:48:04,900 --> 00:48:10,200
constitutes quality as in what 
attributes to reuse to describe 

911
00:48:10,200 --> 00:48:11,700
the quality of data. 
Is it complete? 

912
00:48:11,700 --> 00:48:14,600
This is of Integrity is a 
timeliness, like, is all of the 

913
00:48:14,600 --> 00:48:18,100
above and others So let's define
those and in that definition is 

914
00:48:18,100 --> 00:48:19,700
have subject. 
Matter experts. 

915
00:48:20,100 --> 00:48:23,300
This have domain people who 
actually know their data and how

916
00:48:23,300 --> 00:48:27,200
they can articulate quality 
involved in defining that and 

917
00:48:27,200 --> 00:48:29,800
once that's defined, that's 
automated you. 

918
00:48:29,800 --> 00:48:32,800
Let's put it into the platform 
as a plan for capability. 

919
00:48:33,100 --> 00:48:36,000
The moment, you are 
instantiating it in a product, 

920
00:48:36,000 --> 00:48:39,900
you will get out of the box, a 
library, or some SDK or 

921
00:48:39,900 --> 00:48:42,400
something. 
That gives you the ability to 

922
00:48:42,400 --> 00:48:45,400
now, calculate capture and 
share. 

923
00:48:45,600 --> 00:48:50,700
This quality metrics and then 
you will have observability that

924
00:48:50,700 --> 00:48:53,500
runs across this measure across,
all of his data, paradise, and 

925
00:48:53,500 --> 00:48:56,500
capture, those information 
shares that information also 

926
00:48:56,500 --> 00:48:59,600
validates, whether you are 
meeting the requirements of the 

927
00:48:59,600 --> 00:49:01,300
quality that you've done. 
So that's it. 

928
00:49:01,300 --> 00:49:05,400
Computational part and the 
embedded part that I haven't put

929
00:49:05,400 --> 00:49:09,600
in the title, is that this 
enforcing quality and measuring 

930
00:49:09,600 --> 00:49:13,400
quality becomes an embedded 
concern in every single day 

931
00:49:13,400 --> 00:49:17,000
product, it's not something that
smeared over And added later on,

932
00:49:17,000 --> 00:49:20,500
it's actually from ground up 
built in the data, product 

933
00:49:20,500 --> 00:49:24,200
itself is embedded in there, so 
hopefully that gives a good 

934
00:49:24,200 --> 00:49:28,900
example of achieving I guess. 
Well, oh and and cohesive mesh 

935
00:49:28,900 --> 00:49:33,200
of interconnected data products 
through embedding policies. 

936
00:49:33,200 --> 00:49:37,400
A standard policies in an 
automated fashion in everydays 

937
00:49:37,400 --> 00:49:40,700
product and it have the teams 
that are responsible for 

938
00:49:40,700 --> 00:49:43,500
guaranteeing. 
Those policies involved in this 

939
00:49:43,500 --> 00:49:45,000
defining what these policies 
are. 

940
00:49:45,700 --> 00:49:48,100
Yeah, if we can borrow things 
like from the application 

941
00:49:48,100 --> 00:49:51,700
development rights, we have this
concept as well policy as code. 

942
00:49:51,800 --> 00:49:53,900
There are some tools in 
kubernetes clusters where you 

943
00:49:53,900 --> 00:49:57,200
can embed this kind of policy. 
So before you apply something, 

944
00:49:57,400 --> 00:49:58,900
you will check towards the 
policy. 

945
00:49:58,900 --> 00:50:01,400
And if it's doesn't comply even 
reject. 

946
00:50:01,400 --> 00:50:04,400
So data governance is probably 
one of the least sexy part of 

947
00:50:04,400 --> 00:50:08,100
the data management because they
are things like pii data 

948
00:50:08,100 --> 00:50:10,600
security and maybe should not be
leaked out and it should not be 

949
00:50:10,600 --> 00:50:13,300
exposed maybe things like data 
quality, right? 

950
00:50:13,300 --> 00:50:17,200
How much lagging for example the
data Be and I think all this 

951
00:50:17,200 --> 00:50:20,200
definitely needs to be governed 
because otherwise it's really 

952
00:50:20,200 --> 00:50:24,000
difficult and you mention about 
observability and you use the 

953
00:50:24,000 --> 00:50:27,100
concept from SRE where you have 
also data SLO. 

954
00:50:27,500 --> 00:50:30,300
So maybe if you can touch a 
little bit about this data SLO? 

955
00:50:30,900 --> 00:50:34,600
Yeah absolutely. 
So data products for people to 

956
00:50:34,600 --> 00:50:38,900
trust your data you need to 
share a set of real time. 

957
00:50:38,900 --> 00:50:42,000
I was or at least as real time 
as your data in these 

958
00:50:42,300 --> 00:50:46,400
information to give people trust
that this is a Suitable data. 

959
00:50:46,600 --> 00:50:49,700
So, again, the dimensions of 
that, I think I unpacked it in 

960
00:50:49,700 --> 00:50:51,700
the book and probably don't 
remember all of them are top of 

961
00:50:51,700 --> 00:50:54,100
my head. 
But the dimensions of that all 

962
00:50:54,100 --> 00:50:58,100
around quality, there around 
timeliness there around 

963
00:50:58,100 --> 00:51:01,300
completeness. 
There's a whole set of often in 

964
00:51:01,300 --> 00:51:04,500
the language of potato people. 
These collect metadata is 

965
00:51:04,500 --> 00:51:07,200
language, I don't like and I 
totally use because it's just a 

966
00:51:07,207 --> 00:51:10,600
casual bag of all things. 
But there are classes of really 

967
00:51:10,600 --> 00:51:14,800
information additional data that
you've got to provide for what 

968
00:51:14,800 --> 00:51:18,500
purpose. 
So that the people that want to 

969
00:51:18,500 --> 00:51:23,600
directly self-serve, use the 
product, they can self assess if

970
00:51:23,600 --> 00:51:27,800
this data suits their use case 
or not as an example, the 

971
00:51:27,800 --> 00:51:30,900
distribution of their data. 
So if I'm doing an analysis, 

972
00:51:30,900 --> 00:51:34,000
where trading a machine learning
model for particular, use case 

973
00:51:34,000 --> 00:51:36,800
up, perhaps like to have a very 
I don't know. 

974
00:51:36,800 --> 00:51:41,500
Nice bell curve distribution of 
the data and the samples that I 

975
00:51:41,500 --> 00:51:44,100
can get for training that 
machine learning model, rather 

976
00:51:44,100 --> 00:51:48,000
than by estate us. 
So how by is the data is. 

977
00:51:48,000 --> 00:51:51,800
So these are again SLO is in the
active or more about part up 

978
00:51:51,800 --> 00:51:55,100
time and response time down 
tournaments so on and then the 

979
00:51:55,100 --> 00:51:58,500
data world to the competition of
part of it still has those 

980
00:51:58,500 --> 00:52:01,500
concerns. 
But the data part of it, it has 

981
00:52:01,500 --> 00:52:05,300
a different set of concerns that
defies of usability, Mitch weeks

982
00:52:05,300 --> 00:52:08,200
of a data. 
So thanks so much for explaining

983
00:52:08,200 --> 00:52:10,100
all this. 
It seems like a very crash 

984
00:52:10,100 --> 00:52:13,600
course of data mesh. 
I hope people do study about 

985
00:52:13,600 --> 00:52:16,500
this data may be from reading 
your All your articles are 

986
00:52:16,500 --> 00:52:19,000
watching some of your talks. 
I think it's really an 

987
00:52:19,000 --> 00:52:22,000
eye-opening for those people who
work with traditional data 

988
00:52:22,000 --> 00:52:23,900
management. 
So thank you again for this. 

989
00:52:24,200 --> 00:52:25,900
I have one last question before 
I let you go. 

990
00:52:25,900 --> 00:52:28,900
So normally, I ask these three 
things called three technical 

991
00:52:28,900 --> 00:52:32,000
leadership wisdom. 
Maybe if you can share some of 

992
00:52:32,000 --> 00:52:34,900
your wisdom for us, maybe to 
learn from your journey from 

993
00:52:34,900 --> 00:52:36,700
your experience, or your 
expertise. 

994
00:52:38,100 --> 00:52:42,100
It's a hard woman. 
So maybe just a few things that 

995
00:52:42,100 --> 00:52:44,800
I didn't do as well. 
They can share or things that 

996
00:52:44,800 --> 00:52:48,100
maybe I did. 
Okay, I separate leadership from

997
00:52:48,100 --> 00:52:51,100
management. 
I'm a terrible manager, and you 

998
00:52:51,100 --> 00:52:53,700
don't want me as a manager, but 
maybe I'm going to K be there 

999
00:52:53,700 --> 00:52:56,800
because I believe in the 
mission, I'm a very Mission 

1000
00:52:56,800 --> 00:52:59,200
oriented person. 
So, as a leader, you need to 

1001
00:52:59,200 --> 00:53:02,700
believe in your mission and 
create a mission oriented team 

1002
00:53:02,700 --> 00:53:05,700
and organization and 
continuously through 

1003
00:53:05,700 --> 00:53:09,300
communication, through 
reinforcement of like embodying,

1004
00:53:09,300 --> 00:53:13,400
the bright Behavior to achieve 
that mission reinforce that and 

1005
00:53:13,400 --> 00:53:17,400
remind your teams and keep real.
Lining the team, maybe there are

1006
00:53:17,400 --> 00:53:20,500
different styles of leadership 
but that mission area to 

1007
00:53:20,500 --> 00:53:23,600
Visionary into leadership, 
resonates with me. 

1008
00:53:23,600 --> 00:53:25,500
I love working with people like 
that. 

1009
00:53:25,900 --> 00:53:29,300
And then to get to that mission,
you have two ways of going 

1010
00:53:29,300 --> 00:53:31,400
there. 
You have no way of leaving. 

1011
00:53:31,400 --> 00:53:34,700
May be some casualties behind 
like going in a way that not 

1012
00:53:34,700 --> 00:53:37,200
everyone can catch up and if you
wouldn't take some soldiers 

1013
00:53:37,200 --> 00:53:40,900
along the way, but you want to 
make sure that everybody's alive

1014
00:53:40,900 --> 00:53:44,300
you need to seek about every 
single member of the team, their

1015
00:53:44,300 --> 00:53:48,300
needs their pay. 
Pays their specific hopes and 

1016
00:53:48,300 --> 00:53:51,200
really it's about not only 
having the vision and charging 

1017
00:53:51,200 --> 00:53:54,700
the past but also making sure 
everyone can come along and 

1018
00:53:54,700 --> 00:53:57,700
think Beyond yourself. 
That's probably an area that I 

1019
00:53:57,707 --> 00:54:01,000
need most help with personally 
myself because my mission 

1020
00:54:01,000 --> 00:54:04,400
oriented kind of leadership, 
usually has casualties and the 

1021
00:54:04,400 --> 00:54:08,600
way your people need to be able 
to trust you and believe in you,

1022
00:54:08,800 --> 00:54:12,700
you need to be very self-aware 
in terms of your strengths and 

1023
00:54:12,700 --> 00:54:15,400
your weaknesses where you want 
to delegate. 

1024
00:54:15,900 --> 00:54:18,600
And where you want to actually 
take something on. 

1025
00:54:19,000 --> 00:54:23,800
And if you're a technical 
leader, I personally respect 

1026
00:54:23,800 --> 00:54:28,200
technical leaders that still 
stay close to their craft. 

1027
00:54:28,200 --> 00:54:31,500
They still keep up to date with 
their craft as we all know, to 

1028
00:54:31,500 --> 00:54:35,100
technology moves really fast. 
So you have to find a way to 

1029
00:54:35,100 --> 00:54:38,000
keep yourself relevant and 
up-to-date. 

1030
00:54:38,400 --> 00:54:41,900
And sometimes that means going 
really deep for a moment in 

1031
00:54:41,900 --> 00:54:44,200
time. 
Get your hands dirty and coming 

1032
00:54:44,200 --> 00:54:46,700
back up. 
And of course, As your scope of 

1033
00:54:46,700 --> 00:54:50,400
leadership, Rose, your ability 
to go really deep diminishes 

1034
00:54:50,400 --> 00:54:52,400
because there's time doesn't 
allow for that. 

1035
00:54:52,600 --> 00:54:56,300
So, having carving out space to 
go deep when it's needed, even 

1036
00:54:56,300 --> 00:54:59,200
for a very short period of time.
I've seen some technical leaders

1037
00:54:59,200 --> 00:55:02,700
do that, I admire people who can
strike a balance between the 

1038
00:55:02,700 --> 00:55:06,600
depth and kind of the breadth of
knowledge and relevance of their

1039
00:55:06,600 --> 00:55:08,900
knowledge. 
Well, really beautiful. 

1040
00:55:08,900 --> 00:55:11,500
Thanks for sharing that. 
I think it speaks to some of the

1041
00:55:11,500 --> 00:55:14,400
leaders where they are more 
efficient driven as well rather 

1042
00:55:14,400 --> 00:55:17,100
than managing Well, so thanks 
for sharing that. 

1043
00:55:17,400 --> 00:55:21,100
So, maybe Zama for people to 
learn more from you maybe about 

1044
00:55:21,100 --> 00:55:23,500
data mesh or just to reach out 
and follow up with the 

1045
00:55:23,500 --> 00:55:26,800
discussion piece, their place, 
where they can reach up, well, 

1046
00:55:26,800 --> 00:55:30,600
as it's now really Twitter and 
Linkedin will be the place. 

1047
00:55:30,600 --> 00:55:34,300
So I listened to both channels 
but hopefully soon my company's 

1048
00:55:34,300 --> 00:55:38,500
website will be up and we will 
have jobs and have places for 

1049
00:55:38,500 --> 00:55:41,600
people to reach out directly 
through that, but that's not up 

1050
00:55:41,600 --> 00:55:43,100
yet. 
But when it is, I will let you 

1051
00:55:43,100 --> 00:55:44,700
know and you can share with your
network. 

1052
00:55:45,500 --> 00:55:47,200
Really excited to hear about 
that. 

1053
00:55:47,300 --> 00:55:50,500
So many different data mesh 
Technologies maybe will be 

1054
00:55:50,500 --> 00:55:52,900
coming from that. 
So thanks so much for your time,

1055
00:55:53,100 --> 00:55:55,300
really a pleasure to have this 
discussion with you. 

1056
00:55:55,800 --> 00:55:58,100
It was wonderful to be here and 
we thank you. 

1057
00:56:00,800 --> 00:56:04,100
Thank you for listening to this 
episode and for staying, right 

1058
00:56:04,100 --> 00:56:06,600
until the end if you highly 
enjoyed it. 

1059
00:56:06,700 --> 00:56:09,000
I would appreciate if you share 
it with your friends and 

1060
00:56:09,000 --> 00:56:12,000
colleagues who you think would 
also benefit from listening to 

1061
00:56:12,000 --> 00:56:14,100
this episode. 
And if you are new to the 

1062
00:56:14,100 --> 00:56:17,600
podcast, make Subscribe and 
leave me your valuable review 

1063
00:56:17,700 --> 00:56:20,100
and feedback. 
It helps me a lot in order to 

1064
00:56:20,100 --> 00:56:23,400
grow this podcast better. 
You can also find the full show 

1065
00:56:23,400 --> 00:56:26,500
notes of this conversation on 
the episode page, at Tech 

1066
00:56:26,500 --> 00:56:30,500
Legion, o.f website, including 
the full transcript, interesting

1067
00:56:30,500 --> 00:56:33,600
quotes, and links to the 
resources mention from the 

1068
00:56:33,600 --> 00:56:36,400
conversation. 
And lastly, make sure to 

1069
00:56:36,400 --> 00:56:38,700
subscribe to the show's mailing 
list on package. 

1070
00:56:38,700 --> 00:56:42,300
You know, dot f to get notified 
for any future episodes. 

1071
00:56:42,800 --> 00:56:44,300
Stay tuned for the next 
technology. 

1072
00:56:44,300 --> 00:56:47,200
No episode. 
And until then, goodbye.