 Welcome back to theCUBE's continuing coverage of day one of the Snowflake Summit 22, live from Caesar's Forum in Las Vegas. I'm Lisa Martin, my co-host for the week is Dave Vellante. Dave and I are pleased to welcome Nagaraj Sastri to the program, the vice president of data and analytics at HCO Technologies. Welcome, great to have you. Same here, thank you for inviting me here. Isn't it great to be back in person? Oh, love it. This, the keynote this morning, I don't know if you had a chance to see it, standing room only, there was overflow rooms. People are ready for this, and it was a jam-packed morning of announcements. Absolutely. Talk to us a little bit about the HCL Snowflake partnership, but anybody in the audience who may not be familiar with HCL, give us a little bit of background, vision, mission, differentiation, and then that Snowflake duo. Sure, sure. So let me first start off with talking about HCL. We are a 11.5 billion dollar organization. We have three modes of working. Mode one is everything to do with our infrastructure business and application services and maintenance. Mode two is anything that we do in the cutting edge ecosystem, whether it is cloud, whether it is application modernization, ERPs, SaaS, all of those put together is mode two. Data analytics is part of our mode two culture. The whole ecosystem is called digital services business. And within digital services, one of the arms is data and analytics. We are about a billion dollars in terms of revenues from a data and analytics perspective of the 11 billion that I was talking to you about. And mode three is everything to do with our software services. So we have got our own software products and that's a third of our business. So that's about HCL. So HCL and Snowflake relationship, we are an elite partner with Snowflake. We are one of the fastest growing partners. We achieved the elite level within 18 months of assigning up as a Snowflake partner. We are close to about 50 plus implementations worldwide and about 800 people who are Snowflake professionals within the HCL ecosystem, large customers that we serve. And how long have you been partners? About 18 to 20 months now. Okay, so during the last couple of tumultuous years, why Snowflake? What was it about their vision, their strategy, their leadership that really spoke to HCL as this is a partner for us? So one of the biggest things that we realized probably about four years ago was in terms of, you had all the application databases or RDBMSs, MPPs, the Hadoop ecosystems, which were getting expensive. Not in terms of the cost, but in terms of the processing times, the way the queries were getting created. And we knew that there is something that is going to come. And the people. And the people. And we knew that there will be a hyperscaler that will come. And of course, Azure was already there, AWS was there, Google was just picking it up. And at that point in time, we realized that there will be a cloud data warehouse because we had started reading about Snowflake at that point in time. So fast forward a couple of years after that and we realized that if we have to be in this business, the right way of doing it is by getting partnering, a partnering with the right tooling company. And Snowflake brings that to table. We all know that now. And with what the keynote speakers were also saying, from a 150 member team about five years ago in the conference to about 12,000 people now. So you know that this is the right thing to do and this is the right place to be at. So we devised a methodology in terms of saying that let's get into the partnership. Let's get our resources trained and certified on the Snowflake ecosystem. And let's take a point of view to our customers in terms of how data migrations and transformations have to be done in the Snowflake arena. When you think about your modes, you talked about modes one, two and three. I feel like Snowflake touches on each of those, maybe not so much in the infrastructure and the apps, but although maybe going forward increasingly. So yeah, that's my question is, where do you see Snowflake vectoring into your modes? So it does in both, in the first two modes and mode three also, and I'll give you the reasons why. Mode one is predominantly because you can do application development on cloud, on the data cloud now, which basically means that I can have a COTSA application on Snowflake, eventually, that's the goal. Second is, in mode two, because it is a cloud data warehouse, it fits in exactly because the application data is in Snowflake, I've got my regular data sets within Snowflake, both are talking to each other, there is zero lapse time from a user perspective. And in mode three, the reason why I said mode three was because software as a service or software services and products is because I can, powered by Snowflake, I can implement that. So that's why it cuts across our entire ecosystem. The whole thing is called your digital business, correct? Yes. Right, so that's, this is the next wave of digital business that we're seeing here because it's digital is data, right? That's really what it's about, it's about putting that data to work. So the president of our digital business, Anand Birje, who had done the, who had done a session in the afternoon today, he says the D in the digital is data, he's right. And that's what we are seeing with our customers, large implementations that we do in this ecosystem. There is one other thing that we are focusing very heavily on is industrial solutions or industry-led solutions, like whether it is for healthcare, whether it is for retail or financial services, name a vertical, and we have got our own capabilities around industrialized solutions that fit, that fit certain use cases. So, in thinking about the D in digital is really data, if you think about the operating model for data, it's obviously evolved, you mentioned Hadoop, went to the cloud, all the data went to the cloud. But today it's, you've got an application development model, you got database, which is sort of hardened, and then you've got your data pipeline and your data stack, and that's kind of the operating model. They're sort of siloed to a great degree. How is that operating model changing as a result of data? So, I'll answer it in two parts, part is, if you realize over the years, what used to happen is you had a CIO in an organization or more a CIO, but, and then you had enterprise architecture teams, application development teams, support teams, and so on and so forth. In the last 36 months, if you see, there is an emergence of a new role, which is called the chief data and analytics officer. So, the data and analytics officer is a role that has been created, and the purpose of creating that role is to ensure that organizations will pull out or call out resources within the CIO organizations who are enterprise architects, who are data architects, who are application architects or security architects, and bring them under, into the ecosystem of the data office, from an operating model perspective, so that innovations can be driven, data-driven enterprises could be created, and innovations can come through there. The other part of that is, the use cases get prioritized when you start innovating, and then it is a factory model in terms of how those use cases get built, which is no brainer in my mind at least, but that is how the operating model is coming up from a people perspective. From a technology perspective, also, there is an operating model that is emerging. If you see all the hyperscalers that are there today, Snowflake, with its most latest and greatest announcements, if you see the way the industry is going, is everything will be housed into one ecosystem, and the beauty of this entire thing, and you have to be able to fathom it effectively, because if I'm a multi-cloud kind of an environment, and if I'm on Snowflake, I don't care, why? Because I'm Snowflake, which can work across the multi-clouds, so my data is in one place, effectively. Yeah, it's interesting what you're saying about the Chief Data Officer. The Chief Data Officer, that role, emerged out of the ashes, like a phoenix, of compliance, data quality, and healthcare, and financial services, and government, the highly regulated industries, and then it took a while, but it increasingly became, wow, this is a really front of the board-level role, if you will, data, and now you're seeing it, it is integrated with digital. Absolutely, and there is one other point, if you think about it, the emergence of the Chief Data Officer came in because there were issues associated to data quality, there were issues associated to data cataloging, as to how data is cataloged, and there were issues in terms of trustability of the data. Now the trustability of the data can be in two places, one is data quality, here, bad data, garbage in, garbage out, but then the other aspect of the trustability is in terms of can I do the seven Cs of data quality and say that, okay, I can hallmark this data as platinum, or gold, or silver, or bronze, or un-hallmarked data, and with Snowflake, the advantage is if you have a hallmark data set, that is, I say, a platinum or a gold, and thanks to the virtual warehouse, the same data set gets penetrated across the enterprise. That's the beauty with which it comes, and then of course, the metadata aspect of it, bringing in the technical metadata and the business metadata together for the purpose of creating the data catalogs is another key, a cool thing, and enabled again by Snowflake. What are some of the, when you're in customer conversations, some of the myths or misconceptions that customers historically have typically been making when it comes to creating a data strategy? Some of the misconceptions, and then, what is your recommendation for those folks, since every company these days to be competitive has to be a data company? Yeah, so around data structures, the whole thought process has to be, either doing the past, we used to go with, from source applications, we would gather requirements, then we would figure out what sources are there, do a profiling of the data, and then say, okay, the target data model should be this. Too slow. Too slow, right. Now, fast forward to the digital transformation. There is producers of data, which is basically the applications that are being modernized today are producers of data. They're actually telling you that I'm producing this kind of data. This is the kind of events that I'm producing, and this is my structure. Now the whole deal is, I don't need to figure out what the requirements are. I know what the use case, the application is going to be helping me with, so therefore the entire data model is supported. So, but at the same point in time, the newer generation applications that are getting created are not only getting created in terms of the customer experience, of course that is very critical, but they are also taking into account aspects around metadata, the technical metadata associated within an application. The data quality rules or business rules that are implemented within an application. All of that is getting documented. As a result, the whole timeline from source to profile to model, which used to be X number of days in the past, is X minus at least 20% now, or 30% actually. So that is how the structures, the data structures are coming into a play. Future, a futuristic thought process would be, there will be producers of data and there will be consumers of data. Where is ETL then, or ELT then? There is not going to be any ETL or ELT because a producer is going to say that I'm producing the data for this. A consumer says that okay, I want to consume the data for this purpose. There they meet through an API layer. So where is ETL? Eventually going to go away. Well, and those consumers of data, if you think about the way it works today, the data operating model if you will, the transaction systems and other systems throw off a bunch of exhaust. They get thrown over the fence to the analytic system. They're not operation, the data, the data pipeline, the data systems are not operationalized in a way that they need to be and obviously Snowflake's trying to change that. But that's a big change, please. Yeah, sorry, didn't mean to cut you off. So data operations is a very, very critical aspect and if you think about it holistically, we used to have ETL pipelines, ELT pipelines and then we used to have queries being written on top of Teradata or MPPs and Hadoop and all of that and reporting tools that would have number of reports that were created and certain self-service BI reports into the ecosystem. Now, when you think in terms of a cloud data warehouse, what is happening is the way you are architecting your solution today in terms of data pipelines, those data pipelines are self-manageable or self-healing. You do not need the number of people where there was no documentation in terms of what ETL pipelines were written in the past on certain ETL tools or why something is failing. Nobody knew why something was failing because these are age-old code. But take it forward today. What happens is organizations are migrating from on-prem to cloud and to the cloud data warehouse and the overall cost of ownership is decreasing. The reason is the way we are implementing the data pipelines, the way the data operations are being done in terms of even before a pipeline is kicked or kicked in, then there is a check process to say whether the source application is ready or not ready. So such things, small, small things which are part and parcel of the entire data operations life cycle are taking the center stage. As a result, self-healing mechanisms are coming in and because of those self-healing mechanisms, metrics are being captured. As a result, you know exactly where to focus on and where not to focus on. As a result, the number of resources needed to support gets reduced, cost of ownership is low. Much higher trust, self-service infrastructure, data context in the hands of business users. Data is now more discoverable, it's governed so you can now create data products more quickly. So speed and scale become extremely important. Absolutely. And in fact, one of the things that is changing is the way search is getting implemented. Either doing the past, you created an index and then the data is searchable. But now it is contextual search. Can I contextualize the entire search? Can I create a machine learning algorithm that will actually say that, okay, Nagraj as a persona was looking for this kind of data and then Nagraj as a persona comes back again and looks for some different kind of data. Can the machine learning algorithm go and figure out, okay, what is going on in Nagraj's mind? What is he trying to look at? And then, you know, improve the whole learnability of the entire algorithm. That's how search is going to also take, get into a change kind of a scenario. Excellent, Nagraj. Thank you so much for joining us, talking about data modernization at speed and scale, HCL, what you're doing, what you're doing with Snowflake, and it sounds like incredible power that you're enabling and we're only just scratching the surface. I have a feeling there's a lot more under there that you guys are going to uncover. Sure, so we have a tool or an accelerator. We call it an accelerator in the HCL parlance, but just actually a tool. So when you think about data modernization onto Snowflake, it is predominantly migrating the dataset from your existing ecosystem onto Snowflake, that is one aspect of it. The second aspect of it is the modernization of the ETL or ELT pipelines. The third aspect associated to the data that is there within these ecosystems is the reconciliation. Older application, sorry, older legacy platform, Snowflake. Legacy platform gives me result X, does Snowflake give me result X? That kind of reconciliation has to be done, data reconciliation and testing. And then the third, fourth layer associated is the reporting and visualization. So these four layers are part and parcel of something that we call as advantage migrate. Advantage migrate will convert your Teradata data model into a Snowflake understandable data model automatically. Whether it is Teradata, whether it is Oracle Exadata, Green Plum, Hadoop, you name an ecosystem, we have the mechanism to convert a data model from whatever it is into Snowflake readable understandable data model. The second aspect is the ETL ELT pipeline, whether you want to go from Informatica to DBT or Informatica to something else or data stage to something else, doesn't matter, there is an algorithm or there is a tool which is called the ETL pipeline, we call it GatewaySuit. GatewaySuit actually converts the code, it reads the code that is there on the left hand side which is the legacy code, understands the logic, it reverse engineers and understands the logic, and then what it does is we use that understanding or that logic that has been culled out into Spark code or DBT or any other tool of your choice from a customer standpoint, that's the second layer. Third layer I talked about which is basically data testing, automated data testing and data reconciliation. And the last but not the least is the reporting because older ways of reporting and visualization was of this current day reporting and visualization which is more persona based. The art of visualization is something different in this aspect. Come over to our booth at 2114 and you'll see Advantage Migrate in the works. Advantage Migrate, there you go. Nagaraj, thank you so much for joining us on the program and unpacking HCL and giving us really that technical dissection of what you guys are doing together with Snowflake. We appreciate your time. Thank you, my pleasure. Thank you. For our guests and Dave Vellante, this is Lisa Martin live from the show floor of Snowflake Summit 22. Dave and I will be right back with our final guest of day one in just a minute.