Guest Speaker:
Transcripts
Welcome everybody to Episode 10 of the tech Lyceum series. Our topic for today's podcast is scaling enterprise data management with data fabric and unleashing the value of your business data. Now it's going to be a very exciting episode. And joining us to amp up the excitement is Deepak Gupta practice head of data and analytic transformation services at Birlasoft. And today, we're talking about data management like we've established. So let's go right ahead and give Deepak a warm tech lyceum. Welcome. Thanks so much for joining us today, Deepak.
Follow Us On
Speaker – Deepak - 00:47
Thank you, Neerja. Pleased to be here.
Thank you, Neerja. Pleased to be here.
Speaker – Neerja -00:49
Great to have you, Deepak. Let's dive straight in Data Management is one of the hottest topics in the digital transformation business today. Right?
Q: So, could you talk to us about how data management has changed over the last few years? And where do you think it's headed?
Great to have you, Deepak. Let's dive straight in Data Management is one of the hottest topics in the digital transformation business today. Right?
Q: So, could you talk to us about how data management has changed over the last few years? And where do you think it's headed?
Speaker – Deepak - 01:04
of course. So anyone who has been in data business from last few years, has seen a lot of things change. For example, the introduction of cloud computing, to enterprise IT, the use of SAS and best solutions, IoT, basically the Internet of Things, and now AI and ML. what's noteworthy here, that even the small businesses are using these technologies today. And each of these technologies has not only contributed to the growth of data, but also added to its complexity. Now, there are lot many data sources, hundreds of formats, and even the data architecture has evolved from managing data in legacy applications, on premises databases, to now enterprise level data warehouses, the data Lakes, Lake houses, data Mart's and all in cloud now. At the same time, data management has also become more important. If you look, especially because of security and privacy risk. Any inefficiency in data quality can cause huge losses, and a lot of governance requirements as well. And that's why if you ask the industry is talking about new paradigms like Data Fabric data match,
of course. So anyone who has been in data business from last few years, has seen a lot of things change. For example, the introduction of cloud computing, to enterprise IT, the use of SAS and best solutions, IoT, basically the Internet of Things, and now AI and ML. what's noteworthy here, that even the small businesses are using these technologies today. And each of these technologies has not only contributed to the growth of data, but also added to its complexity. Now, there are lot many data sources, hundreds of formats, and even the data architecture has evolved from managing data in legacy applications, on premises databases, to now enterprise level data warehouses, the data Lakes, Lake houses, data Mart's and all in cloud now. At the same time, data management has also become more important. If you look, especially because of security and privacy risk. Any inefficiency in data quality can cause huge losses, and a lot of governance requirements as well. And that's why if you ask the industry is talking about new paradigms like Data Fabric data match,
Speaker – Neerja - 02:17
Right,
Q: Let's also maybe understand what does a Data Fabric mean? Could you enlighten us?
Right,
Q: Let's also maybe understand what does a Data Fabric mean? Could you enlighten us?
Speaker – Deepak - 02:24
Firstly, this Data Fabric is like a word everybody's trying to understand. It's not a it's just a map data management design. It's not a tool, or technology, first of all, right facilitates pasture, high quality, Integrated Data View, to all consumers abstract abstracting the data complexity, which is actually caused because of heterogeneous data sources, different deployment approaches, etc. So what data factory actually does, it makes use of metadata, the semantics, which came out from all the different sources, build a centralized data asset, with actually act as a intelligent ox orchestration engine to recommend like flexible data integration pipelines, and in fact, bring the automation in the entire data availability and use it.
Firstly, this Data Fabric is like a word everybody's trying to understand. It's not a it's just a map data management design. It's not a tool, or technology, first of all, right facilitates pasture, high quality, Integrated Data View, to all consumers abstract abstracting the data complexity, which is actually caused because of heterogeneous data sources, different deployment approaches, etc. So what data factory actually does, it makes use of metadata, the semantics, which came out from all the different sources, build a centralized data asset, with actually act as a intelligent ox orchestration engine to recommend like flexible data integration pipelines, and in fact, bring the automation in the entire data availability and use it.
Speaker – Neerja - 03:15
All right, so Deepak you mentioned that Data Fabric is a data management design pattern, right?
Q: How is this implemented? And will all organizations implement it in the same way?
All right, so Deepak you mentioned that Data Fabric is a data management design pattern, right?
Q: How is this implemented? And will all organizations implement it in the same way?
Speaker – Deepak - 03:26
Oh, Excellent question. First of all, data doesn't refer to a single architecture blueprint. What does it mean? Like you said, right, it's a pattern, which has some essential components, which are deployed depending upon what business need and what is the maturity of an organization.
For instance, you have an augmented data cataloging layer, which manage all data assets, then there is a meta data layer, which will stitch active and passive metadata, which is coming from all sources. And similarly, there will be a knowledge graph layer with semantic analytics as a recommending engine. So there are multiple layers which are deployed based on the business needs. Okay, this
Oh, Excellent question. First of all, data doesn't refer to a single architecture blueprint. What does it mean? Like you said, right, it's a pattern, which has some essential components, which are deployed depending upon what business need and what is the maturity of an organization.
For instance, you have an augmented data cataloging layer, which manage all data assets, then there is a meta data layer, which will stitch active and passive metadata, which is coming from all sources. And similarly, there will be a knowledge graph layer with semantic analytics as a recommending engine. So there are multiple layers which are deployed based on the business needs. Okay, this
Speaker – Neerja - 04:11
Sounds interesting. Deepak you also mentioned data mesh A while back,
Q:How is that different from a data fabric?
Sounds interesting. Deepak you also mentioned data mesh A while back,
Q:How is that different from a data fabric?
Speaker – Deepak - 04:18
Excellent question again. Data mesh is a very different concept, where you aim for decentralization of data management based on domains with a data mesh data assets are managed like a product. And each of these data products are managed by business domains. While Data Fabric is more technology centric design, versus data mesh is more people or process focused. You see, you know, data mesh is focused on organizational change. It brings greater autonomy to business domains, and encourage them to build business Aligned Data Products.
Excellent question again. Data mesh is a very different concept, where you aim for decentralization of data management based on domains with a data mesh data assets are managed like a product. And each of these data products are managed by business domains. While Data Fabric is more technology centric design, versus data mesh is more people or process focused. You see, you know, data mesh is focused on organizational change. It brings greater autonomy to business domains, and encourage them to build business Aligned Data Products.
Speaker – Neerja - 04:57
Q: So can we draw a comparison here is one better than the other.
Q: So can we draw a comparison here is one better than the other.
Speaker – Deepak - 05:00
Honestly, not really. Data mash and Data Fabric solves the problem of data management though, but in two different ways. When you use a data mash, you effectively distribute the responsibility of data management to your domain teams. It's more like a decentralization approach, which brings more autonomy to domain teams. On the other hand, Data Fabric centralizes data management. But in addition to that, also a lot of other stuff. Like, for example, bringing simplicity, standardization to data integration, storage, pipeline processing, automation, etc. Right. That's why when you have a data mash, you don't have a single point of failure. it may bring better agility, but at the cost of standardization. But with a Data Fabric, you get consistent data governance. In fact, you can implement data quality, you can implement security, stories standards, all across your data assets. In fact, you can reduce data and and storage and processing and streamline the entire analytics layer, including reporting, although it may slow down domain specific innovations. And it can slow down experimentation, because your team will not have the autonomy to explore new technologies. So they are, I would say, there are two different ways to solve the same problem, and which has pros and cons of each other.
Honestly, not really. Data mash and Data Fabric solves the problem of data management though, but in two different ways. When you use a data mash, you effectively distribute the responsibility of data management to your domain teams. It's more like a decentralization approach, which brings more autonomy to domain teams. On the other hand, Data Fabric centralizes data management. But in addition to that, also a lot of other stuff. Like, for example, bringing simplicity, standardization to data integration, storage, pipeline processing, automation, etc. Right. That's why when you have a data mash, you don't have a single point of failure. it may bring better agility, but at the cost of standardization. But with a Data Fabric, you get consistent data governance. In fact, you can implement data quality, you can implement security, stories standards, all across your data assets. In fact, you can reduce data and and storage and processing and streamline the entire analytics layer, including reporting, although it may slow down domain specific innovations. And it can slow down experimentation, because your team will not have the autonomy to explore new technologies. So they are, I would say, there are two different ways to solve the same problem, and which has pros and cons of each other.
Speaker – Neerja - 06:24
Right. That's quite a comprehensive comparison you've given us there, Deepak and let me ask you this in your work with clients, right?
Q: What are some of the key benefits that our data fabric has delivered?
Right. That's quite a comprehensive comparison you've given us there, Deepak and let me ask you this in your work with clients, right?
Q: What are some of the key benefits that our data fabric has delivered?
Speaker – Deepak - 06:37
Well, we have worked with a lot of clients across many industries, where they're disrupting their sector, either to cutting edge data analytics use cases, we have customers who are building customer custom data platforms like fraud, detections, preventive maintenance, risk modeling, return to Office, etc. But one observation that has been consistent. A is data management is a crucial aspect to bring method to integrate the data. And what you find out specifically with fabrics, implementation, organizations were able to reduce human effort by almost 50%. Through all these different automation mechanism and orchestration styles, and more importantly, the data adoption and utilization increased by 400%. Although such mattresses are subject to context of business, but we are seeing tremendous improvement in those areas.
Well, we have worked with a lot of clients across many industries, where they're disrupting their sector, either to cutting edge data analytics use cases, we have customers who are building customer custom data platforms like fraud, detections, preventive maintenance, risk modeling, return to Office, etc. But one observation that has been consistent. A is data management is a crucial aspect to bring method to integrate the data. And what you find out specifically with fabrics, implementation, organizations were able to reduce human effort by almost 50%. Through all these different automation mechanism and orchestration styles, and more importantly, the data adoption and utilization increased by 400%. Although such mattresses are subject to context of business, but we are seeing tremendous improvement in those areas.
Speaker – Neerja - 07:32
That's great to hear. And I'm sure our audience has found some very compelling insights on data management in this episode. So thank you so much for joining us, Deepak and allowing us to share in your wisdom.
That's great to hear. And I'm sure our audience has found some very compelling insights on data management in this episode. So thank you so much for joining us, Deepak and allowing us to share in your wisdom.
Speaker – Deepak - 07:45
Oh, my pleasure, joining you Neerja. And just to give you an idea, we at Birlasoft have very strong data analytics, competency. We support end to end data value chain, right from data platform modernization up you know, data engineering and integrations, data visualizations, have lot of focus on AI machine learning data sciences, and we also have very strong unit around master data management, data governance.
Oh, my pleasure, joining you Neerja. And just to give you an idea, we at Birlasoft have very strong data analytics, competency. We support end to end data value chain, right from data platform modernization up you know, data engineering and integrations, data visualizations, have lot of focus on AI machine learning data sciences, and we also have very strong unit around master data management, data governance.
Speaker – Neerja - 08:15
That's fantastic Deepak. Thank you so much.
That's fantastic Deepak. Thank you so much.
Speaker – Deepak – 08:16
Thank you.
Thank you.
You were listening to tech Lyceum, a podcast by Birlasoft
Other related Tech Lyceum episodes
