William: Welcome! Data Decoded Episode #3, an IBM podcast series dedicated to demystifying the world of data from data lakes to master data management to date science, big data and everything in between. This is the podcast for the data professional and all those who understand they are in the business of data. I217;m your host, William McKnight, President of McKnight Consulting Group – now celebrating twenty years of consulting this year. Time flies, let me tell you.
Today, I am joined by Soma Shekar Naganna, Software Architect of the Analytics platform at IBM. I am fascinated by what Soma is doing and he’s eager to share, so that makes a perfect combination for you. Now, I love it when worlds collide when subspaces within data come together and that’s exactly what we have for you today. Our last podcast was all about blockchain and Soma has been busy applying blockchain to the discipline of master data management. I’m going to let Soma explain, but welcome Soma!
Soma: Hi William, thank you for the interaction.
William: Well let’s get right into it – I know Hyperledger is part of your innovative solution there. Can we start out with some insights into Hyperledger?
Soma: Hyperledger is an open source project. It’s basically building blocks for any blockchain initiative so you can convert a private blockchain network into a public blockchain network and IBM is one of the major companies to back this open source project. In 2015, IBM nearly wrote 40,000 lines of code to the Hyperledger.
William: So [if] somebody out there were interested in blockchain, they could get Hyperledger as their foundation?
Soma: Absolutely. This is available as open source. Anybody can download it and anybody can I start using it.
William: So what are some of the things that IBM has done to add value to the open source project?
William: Well those are always good things to do with the open source projects to make them enterprise-ready there. Speaking of the Hyperledger application inside an enterprise, I’m curious as a data professional, what each party in the blockchain is required to do so we know it’s a trusted mechanism for sharing information across a chain of parties, but what does each party need to do in terms of hardware and software to be part of the blockchain?
Soma: First of all, interested parties need to come to an agreement on what do they really want to share on blockchain. So once they come into an agreement, what are they really sharing may be a customer information that they’re sharing or it could be a trade agreement, a sort of an agreement that is shared across multiple parties one-in-one. That turns out to be a smart contract which is nothing but a programming element that gets executed on the Hyperledger network.
Now coming back to what is a network from a software perspective, it’s a container in itself which has abilities of blockchain. Basically, there are nodes which get instantiated at multi-part level. Those nodes interact across the network.
From a hardware perspective, there are multiple things that you can do. You might know that IBM has one small grain of computing node you can post your blockchain or Hyperledger. Since it is an open source you can really have your commodity hardware instantiated. From a software requirement respect, you can use a specialized hardware like what IBM has invented, you can use a commodity hardware.
William: Okay well that’s a great overview for data professionals out there. Now let’s Segway a little bit to master data management because you’re going to apply blockchain to master data management – I should say IBM has applied this blockchain concept to master data management. And for those who don’t know about master data management out there I do a lot of consulting in this space and it really means a lot of things to a lot of different people.
I find a lot of people leave value on the table even when they’re deploying a master data management but we’re trying to get to a single version of a set of data that is widely optical across the enterprise so it might be your customer list, your product list, your site list except for things like this in addition to good old reference data so we build it and sometimes we build that by drawing it in or integrating it from different places where it exists today already in the architecture and sometimes work flow issues to build that data.
Always data quality applies to it and that data is immediately distributed to all the subscribers to that data in the environment. Hopefully there’s many up there. So that’s just a quick nutshell. it’s also great at hierarchy management and data quality, as I mentioned, things like this. So a lot of organizations now are taking up master data management. It took a little while, I think, for it to hit its stride but it certainly has now and organizations are seeing the value of building this master data once and using it many times. That’s exactly what master data management is all about. So, I’d like to know, Soma, what gave you the idea that blockchain would be useful to master data management.
Soma: That’s the million dollar question if you ask me. I’ve been at IBM for more than 12 years and I’ve done all sorts of master data management throughout my career, developing aspects of master data management. What struck me was that Hyperledger is nothing but a multi-party business network with the data you share, which sort of rang a bell for me that some sort of thing that we do in MDM, we either use the reference data which is shared across multiple parties or we use master data which is also shared across multiple parties. So I quickly realized that, can’t we do something on blockchain because all of our data sharing….if somebody would share the customer profile, can I do a data quality aspect like matching and linking. Can I serve those records back to the calling business user. That’s where the idea of master data management on blockchain struck me.
It took a series of validation I had to undergo – it could be an internal or external. I work with various customers, some large banks or garment agencies, health care domain. I kind of figured out that it really has some value that can be leveraged by putting MDM on it.
William: Absolutely, like I mentioned, master data is data that is widely applicable across the enterprise so if you’re going to distribute that data, the blockchain makes a lot of sense and the more I think about it since we last talked, I think that this could really revolutionize master data management I think it’s really a strong and a way to go. Kind of hard to believe blockchain has been around for twelve years though and you’ve been working on it for all of those twelve years – that is great!
Now Soma, you touched on some of the potential customer situations. For this, you mentioned banks and healthcare. Can you drill in a little bit though on one or more of these customer situations to make this a little more concrete?
Soma: Okay, so let me take an example which is widely talked about and maybe is very simple to understand. Today, we have a healthcare information exchanges which is nothing but a data consolidation where an organization with multiple [parties] – such as hospitals, clinical care, laboratories – would collaborate together and share the patient information. Today, what we have is to centralize them and create an information exchange.
Once you really centralize the data, then that becomes a bottleneck in specifically in terms of data quality aspects, right. The data quality aspects of all the agreed parties have to be applied, are in some cases, some of the organization needs to downgrade the data quality aspect. So what I realized was if we can really eliminate a centralized system [and] everybody shares their data across the network not having this healthcare information – which could be non-profit organization, public organization, or a private organization.
I see that there is really no one need of that with the blockchain and then you can really share the data all across them and you can have your own data quality aspects because common data you shared across and then you build your own data quality aspects. The entities are the data with respect to their data quality so isn’t there something disrupting thing to do – so that’s where we should be working.
William: Well, that’s fascinating and indeed if you’re distributing the data widely across the enterprise and beyond really, you really need a lot of good data quality in there and master data management is really known as the discipline that provides a data quality so this is really great to see. As a matter of fact, the term killer app comes to mind with me and I don’t know if MDM is the killer app for blockchain or if it’s the other way around, but how do you see it Soma – is MDM going to be the killer app for blockchain?
Soma: So I would really say it doesn’t, yes, because there are a lot of aspects like today if you want to centralize the data, you still have to do an EPO in process once you extract the data, once you import that. I mean all the guys who have done MDM will know about the pain points that they undergo and about the time and money that we need to move the data across. If you really think about it as gone out of a system, it’s a big wound for people who are doing MDM. I believe that’s one of the reason I believe this is a killer app.
William: Great, well when is this going to be available and how is it going to be available. Is it going to be inside the core IBM MDM product?
Soma: This is going to be an all foot on top of Hyperledger. We are still trying to finalize the modalities of how we can really offer it. But the current state is data, which is available for any of the existing customers of IBM MDM. They can really use and try out this.
William: So what is the best way for somebody who’s interested in this to get in touch with you or IBM to get some more information?
Soma: It’s me – I will share my email ID. So anybody reaching out to me in the case of blockchain and MDM, I’ll be more than happy to take them through and most importantly, I believe that this is one of the killer app kind of stuff that I really want people to participate and also be part of this journey – so that’s my take on this.
William: Absolutely, so if everybody would think that’s out there that’s doing it master data management using the blockchain for distribution and everybody out there doing blockchain, I can think about the data, the master data, that they really do want to distribute on the blockchain – I think that these two concepts can come together really well.
Thank you very much Soma for joining us for this episode of the Data Decoded podcast. You can find this podcast and more at ibmbigdatahub.com, also on iTunes of course. You can reach me through my company at McKnight Consulting Group – mcknightcg.com. Until next time – thanks for listening to Data Decoded.