Vincent Pineau, General Manager Americas, Talend
Bob Zurek: Hi, welcome to the second in the series of EnterpriseDB podcasts covering topics including Postgres Plus, PostgreSQL, partners who support Postgres Plus, and the state of the open source industry. I'm Bob Zurek, CTO of EnterpriseDB and your host for this podcast. Today we would like to welcome Vincent Pineau, who is the general manager of North America for Talend. As many of you know, Talend is a market-leading open source data integration company. I'm really pleased to have Vincent participate in today's podcast, and I'm very excited to have the chance to speak with him about the state of affairs within the Talend community, the Talend company, and also about the current release of Talend's solutions and what might be coming in the future. Good day, Vincent! Great to have you on the podcast today. Can you give me a brief background on Talend and your efforts in the market?
Vincent Pineau: Certainly, Bob. Hello, and thank you for giving me the opportunity to participate in your great podcast series. Talend is actually a great story, like no other. It's a company that was started based on the observation that there was an underserved part of the data integration market, and that really sets the tone as far as what Talend is all about. The founders were data integration experts already, and they observed that the tools that they had to propose to their prospects were not really hitting it where they should have, being sometimes too cumbersome, or too expensive, or not appropriate for the task at hand. They thought that there was a market opportunity, and that's how Talend came to mind, and that's what drove Talend to what it is today: a tool that aims to democratize data integration, making it available to companies of all sizes and for any kind of data integration needs.
BZ: When we talk about data integration, that encompasses a lot of capabilities. How does Talend describe data integration? What do you see as the core features of a data integration solutions?
VP: If you look at data integration, you can basically slice it into two domains. The first is the one that deals with analytical data integration; that is the practice of ETL-ing (extracting, transforming, and loading) data into a data model or a data warehouse. That part of data integration is well understood, and is where most of the players have been focusing. Another part of data integration entails anything that is operational, which really includes everything else from your data migration needs . you may need to move data from an old system to a new one, or you're moving from an antiquated CRM solution to a newer and better CRM. How do you do this? Up until we came into the picture, there was very little possibility other than custom coding. That also holds true for synchronization of your application; if you have a CRM system or a financial system, that sit on their own respective applications, it's highly likely that you're going to benefit a lot from having the respective customer information be synchronized between one and the other. And that holds true for other functions like application upgrades, so that market really is a huge part of the market. Actually an IDC research a couple of years ago suggests that the size of the operational data integration market is just about 5X the size of the analytical one, with a global expenditure being respectively $11 billion and $2 billion.
BZ: That's a huge market, and, as we know, information continues to grow in an organization, and frequently when projects get started they end up creating silos of data and trying to get all of that data into, for example, a data warehouse. Probably one of the key features of Talend is trying to get all of that data together into a repository for analytics. I understand too that a lot of companies, in their efforts to reduce costs, try to do as much "instance consolidation" . for example, they might have three, five, six instances of SAP and they'll try to bring that down to two or three. Would you say that data integration plays a key role in this instance consolidation effort for an organization?
VP: Most certainly, yes. This would fall in the second domain of the integration, which I defined as "operational data integration." We're seeing that as a key driver to data integration.
BZ: If we step back a little bit and shine a spotlight on the Talend solution, what would you say are some of the cool features that really differentiate Talend from the commercial vendors that are offering data integration? Obviously you're a leader in this space, so you must have gotten some great feedback from your community and your fast-growing customer base. What do they like?
VP: It's always funny to think of a data integration solution as "cool," and the fact of the matter is, no matter how highly I think of Talend, we are plumbing, we are the pipes and nuts and bolts of your data, so I don't know that "cool" defines it most appropriately. But Talend definitely is a highly-differentiated solution, and, historically speaking, we came about with a solution thinking that there was a void. So, to highlight some of those key differentiators, I'd say that we are a co-generator, which to some might be a surprise, but inherently carries great differences and great advantages. As a co-generator, once you define your job in a highly-protected, easy drag-and-drop environment, you're going to be able to run that job on virtually any platform, from Windows to Linux as well as Unix, Mac OSX, you name it. That is something that is extremely important to our community and very helpful. Also, being the open-source solution that we are, we decided to focus on versatility and extensibility, and one of the ways you see that in your solution is that we have over 250 connectors to reach out to the data in most of the amazing places that it could be stored, ranging from your databases to your web services, RSS feeds . you name it, we probably have it. What's more important is, if we don't have it, it is extremely likely that the community came up with it and made it available back to the rest of the community through our ecosystem, which is a place where users and customers continue to add new features of Talend by contributing content.
BZ: How are things going within the community these days?
VP: We are very thankful to have the community we have. We're seeing a lot of traction; we have an aggregated number of downloads of over 1.7 million since we came to life. We have an extremely active community, not only by the contributions that I alluded to a minute ago, but in the shape of providing feedback or being extremely active in the forums. So we are really excited about having the community that we have, and seeing it grow at the pace at which it's growing, which is, on average, one new download per minute, one new paying customer per day. Everything is pretty much green light, fast lane, and we are very excited about the market that we've entered and the open source approach to that market, which seems to be a good combination.
BZ: As you indicated earlier about the connectors, I know Talend connects to Postgres Plus and PostgreSQL and Advanced Server, and it's great to have you a partner and ensure that our customer base has the opportunity to take advantage of your data integration product. I also understand that you've made available a new data profiling tool; how's that going?
VP: This is something that's been in the works for awhile, because, at the end of the day, when you think of data integration, you can think of three main activities. I like to think of data transportation, data transformation, and data profiling. When you bring data from lots of different sources, or when you try to match data, the quality of the underlying data plays a key role in how complete your solution is going to be, how successful you're going to be bringing that data together. Being the first provider to offer an open source data profiler, we feel that we fill the void and address a need that everybody has without necessarily understanding it. Data profiling ability allows one to connect to their database and do introspection, very easily identifying which tables need to be paid attention to, which columns and whatnot, and the tools can automatically tell you the accuracy of the underlying data by running some typical statistics like "count the nulls." In Talend's view, data profiling, and data cleansing, is an integral part of data integration, and it's extremely important to have that in your data integration solution.
BZ: Data quality is a very important part of the data integration process, and I hear some rumors about Talend approaching that area too as part of their overall strategy.
VP: I can just say that you're well-informed. I'll just say that, when you add data profiling with data cleansing, you can think that data quality is probably going to pretty well drop. So, more to come on that.
BZ: Any final comments about your solution?
VP: It's available on our web site at www.talend.com. We let our users install the tool; we are always available to help them through the experience. They can figure it out, take a test drive, and see what value they can drive of the tool.
BZ: Thank you for taking the time to spend with me today on this podcast. I look forward to hearing more about Talend in the future. Thanks again, Vincent, for all of your time, and have a good day.
