Gaurav Dhillon knows a thing or two about integration. In his twenties, he co-founded Informatica and helped thousands of enterprises deal with the challenges of application and data integration in the client-server world. Now, as CEO of San Mateo, California-based SnapLogic, Dhillon is tackling the integration challenges IT shops face in the new world of cloud.
In this installment of the IDG CEO Interview Series, Dhillon spoke with Chief Content Officer John Gallant about how SnapLogic unites cloud apps and legacy systems, and how customers are using the company's tools to ease the move to next-gen data and analytics platforms. He also discussed why data explosion of the internet of things will be a boon for SnapLogic.
Why was SnapLogic founded and what has been the path to date?
The short version -- not the barefoot-in-the-snow-uphill-both-ways version -- is that we put together a project about 10 years ago to investigate whether, using modern web technologies, we could transform the industry of integration. Once we came to that reality that we could, I joined as CEO, we brought in some venture capital, and we've been a corporation since about 2010. It took a few years to get the science right and we got the engineering going on the back of that.
SnapLogic CEO Gaurav Dhillon
Some of us in this company come from backgrounds in integration in the '90s. I built Informatica as its founder and CEO. I was there for about 12 years, built it into a couple-hundred-million-dollar market-leading company. It was a good run for a 25-year-old entrepreneur.
That said, what I felt was that while we had done very well with the transition from mainframes to client server, we sort of missed the boat on what the worldwide web was doing in business. You could use browsers, load balancers, edge caching and move information around the enterprise. People would no longer use applications, but would use cloud computing, a browser, and some sort of SaaS website to do their sales management, human capital management, help desk functions, and so on and so on. There is a need and therefore an opportunity to provide a new kind of technology to connect all these things that are now moving to the cloud with many of the things that are going to stay on premise for the longest time.
I was driven by the renowned thinker at Harvard Business School, Ted Levitt, who famously wrote a premise on marketing myopia. I'm a huge fan of Ted's writing and his thinking. What became obvious to me in my time after Informatica is that we had been thinking about the business narrowly because of the technologies we had on hand in the last century. If you apply that premise to this industry, you realize that Levitt was right. His famous quote was, "Railroads were never in the railroad business, they were in the transportation business."
Our realization at SnapLogic was that the legacy integration business, including my youthful misdemeanors, had been narrowly focused on railroads. At companies like Informatica we did things like ETL (Extract Transform Load). Across town, the founding team at Tibco did things like EAI (Enterprise Application Integration). They did app integration and enterprise software buses and so on but the reality is we were all in the data transportation business. It just is that we didn't have the technology to be able to do that using a single platform. At the risk of now giving you the long version, that has been the journey and SnapLogic has been building one platform to be used for multiple problems that exist in the modern enterprise.
Take someone like Denny's. They provide food, they do a wonderful job. The Grand Slam breakfast is a fun thing to do. They are going to the cloud. They are bringing in new cloud applications. How do they connect them to all the stores and all the other things they have? They Googled around, they talked to Gartner, and they ended up picking SnapLogic to connect the modern applications to the legacy [applications] they have.
What are the specific challenges that you're trying to address?
We address three markets with our platform. The first and certainly the biggest today is enterprises moving to the cloud. We call it the cloudification of the enterprise. Whether it's AstraZeneca or Denny's or Box or Uber or Amgen -- which recently became a customer -- these folks are moving to the cloud for new applications. They're buying things like Workday and Salesforce, and when they buy these new enterprise SaaS applications they need to connect them with the rest of the technology they have.
For example, if you want to onboard a new employee at a modern enterprise you realize that half your processing is being done in the cloud and half on premise. You might require accounts for them on LDAP. You might require connectivity into your financial system and, certainly, for using the human capital management system from Workday. Their salary information, their background information, all the HR data is in the cloud. To do that onboarding of a new employee and to make sure that raises and salaries and everything is reflected and their photograph shows up in the email directory, there is a good old-fashioned integration across cloud and on-premise that has to be done. That's one end of it.
The other end of it is more sophisticated things like you bring in Salesforce, you're moving your CRM into the cloud, and now when you go from a quote-to-cash when you close an order, you ring a big bell or you tweet about it, I guess, these days. Nobody rings a bell. Then you want to have that process generate a purchase order and generate that information in your financial system be it SAP or Oracle or even Workday Financial. How do you conclude that loop?
That's the sort of thing that SnapLogic does and does really well. The principal problem that we solve is helping enterprises move to the cloud at a very fast clip.
In addition, and I think over time it might become a bigger use of our platform, is we help enterprises move to the new data platforms. Capital One is, to my mind, the pioneering company in data science. They issue credit cards to people whom others might deny. By using data science they're able to figure out how to issue them credit cards and create a fantastic business for themselves and help people function in a society where without a credit card you might as well not have any money.
When companies like Capital One are moving away from legacy data warehousing environments into modern data platforms using Hadoop, they look to companies like SnapLogic to help them move their data. We are one platform that is helping enterprises move to the cloud as well as move their data for analytics.
Now what we're also starting to see is that there are projects on the horizon across industrial and even consumer-facing companies -- particularly health care -- around the internet of things. If you think about the legacy of data warehousing, a lot of it came from the bar code scan. You went to the grocery store, you bought a bar of soap, the checkout clerk zapped it with a laser, it automatically added it into your shopping basket, and boom, you pay for it and you're done. That information is hugely valuable for retailers to understand what's selling more, Coke or Pepsi, how to restock the shelf and how to do couponing and planning and demand generation, demand planning and price optimization. All that industry, which is tens of billions in size, came from the bar code scan.
Imagine a world where everything is an internet-connected sensor of some type that is emitting data that is going to make bar code scans look like a pebble on the beach when we look out at the ocean. That ocean of data that is coming our way with the internet of things is going to be a huge market potential for our company. We'll help modern enterprises make better products using data and make money by selling some of those analytics back to the customers.
In my conversations with some of the leading cloud application providers I've actually wondered about this emerging integration challenge as companies go to these best-of-breed cloud providers, whether it's HR or CRM or finance or whatever the tool. Can you put that in perspective for CIOs and other senior IT executives? What are the looming integration issues here that people should focus on more?
You're right. Integration is a lot like oxygen. You don't miss it until it's gone. I think the problem that is looming is that we are in an age where every company is going to have its own technology fabric. With the growth in SaaS, the growth in cloud computing, more and more choices at reasonable cost are available to the technology executive. He or she can build their own ERP and assemble it based on what is best for them if they're an insurance business or a pharmaceutical business or health care or consumer packaged goods. Depending on their industry and their particular position in that industry, they might make different choices.
This concept of business process reengineering and one-size-fits-all solves all problems, is a last-century, quaint idea that is not holding up in the modern enterprise. Businesspeople can use these things in marketing, their children are using them on iPads, and they come into the office and say: Why doesn't it work like this?
We should really be thinking about how do we speed up the way in which a traditional technology executive can help his business colleagues do their work better, whether they're in human resources or in marketing or in sales or manufacturing. How we can help them do their jobs better is by giving them their own technology and put that together so that the thread that creates this fabric is integration. That's how you pick and choose.
You might pick financial system A and you might pick human capital management system B or you might pick something in one area, something else from another area. You might use Hadoop on premise for your manufacturing people because all the machine data is on premise, and you might choose to put your marketing in the cloud with Microsoft HDInsight product in Azure.
These choices should be very, very easy. It should be like provisioning something rather than getting into a hardware, air-conditioning type of procurement cycle. Integration makes that possible. Integration is the glue, the thread that brings this technology fabric together for the enterprise, so they can have their own solution and they can work at an optimal level and they don't have to buy the same thing as everybody else. It's not just beige. That was the '90s. In business process reengineering, everything is beige.
Now, in this millennium we have so many good choices. Putting them together, making them work as one is, I think, the magic that integration brings into the modern enterprise.
What is the name of the product and how is it delivered? How do you build and support this?
It's a true cloud product. There is a lot of fake cloud. You download this and download that and oh, here's some cloud orchestration. To me, that is not a cloud product. If you can Google a product name and version and you find 16 versions on GitHub, that's not really a cloud product. That is cloud washing.
We've taken an approach of having a true platform as a service. We built the company on that. We took a lot of capital, a lot of time to do it. We have a very simple, easy-to-use, user experience because in this day and age somebody using a smartphone is intolerant of legacy UX.
The way this works is we layer out the simple UX with grids of SnapLogic processors. We call them nodes and we run them in a network configuration. We're using the Latin word "plex" as our descriptor for it. "Plex" is Latin for "network." You have Snaplexes, which do the runtime processing of our data integration pipelines. You might have them in the cloud, called a Cloudplex. You might have a Groundplex, you may have a Hadooplex, you may want to run it on Spark. Here I am in my conference room looking at one of our graduate student's project this summer where he and his colleague put together a Snaplex running on a Raspberry Pi, so we have a Raspberryplex. Then you have to now wire it all together.
You need Snaps, which are prebuilt intelligent connectors. We pioneered the concept of containerization of data in the modern enterprise through Snaps. It's not something we invented. We've just taken that idea to an application level and come up with the concept of Snaps, which are standard-sized containers for connecting everything from Twitter to Workday to SAP to Salesforce to you name it. We have over 400 Snaps.
That's in a sense our layer cake. There's a simple UX, Snaplexes that can run on the ground, in the cloud, on Hadoop, Spark, or whatever you have, and there are 400-plus Snaps that are on a canvas. You can drag them in and Snap in something that otherwise would be a lot of complexity. It might take thousands of lines of Java code to do what we can do on a drag-and-drop interface.
SnapLogic
SnapLogic architecture components
The product is called the SnapLogic Elastic Integration Platform. It is available through a direct sales model and it's available globally. We expanded internationally in a very nice way. We just brought on our first Australian customers. We're having a good summer here and have lots of folks all over the world using it, some very big companies in Europe.