David Inbar, Pervasive Software's director of Worldwide Marketing for Integration and International Channels, explains the most overlooked problem when it comes to SaaS integration, as well as what's new for Pervasive, in this interview with IT Business Edge's Loraine Lawson.
Lawson: Pervasive started out as an on-premise solution, correct?
Inbar: That’s correct. The roots of the original integration company, Data Junction, were actually on-premise dealing primarily with legacy data - mainframe, COBOL, all those lovely things with copy books that matched or didn’t match and so on. The vision then was all this machine-readable data out there, it’s really tough for humans to understand it and be able to quickly move it around into different formats. We’ve been through several iterations since then. We’re probably into our fourth re-architecting. As we’ve grown, the number of connectors moved from desktop to client server, from client server to server-based distributed, to cloud-based.
Lawson: Where are you seeing the most growth?
Inbar: Certainly I would say in terms of growth point, integration tends to go where systems are changing or expanding. So it’s very natural for a company like us to see growth around SaaS in particular.
We have some customers that are moving themselves entirely to SaaS-based applications and cloud. Those are interesting because the integration challenge doesn’t go away, it just changes. The change they're experiencing is that, in some ways, it’s easier to change applications if you're subscribing to a SaaS-based offering or set of offerings because, financially, you can turn it on and off faster, you don’t have an on-premise investment in hardware and in IT stuff and so on. So on the face of it, you could switch things on and off and jump onto different vendors.
In practice, you still have to take care of the fact that those applications integrate with what you’ve already got or with other SaaS apps, and so what's your strategy in handling that? It could be very disruptive if you switch a CRM system or a help desk tracking system and the data suddenly isn’t flowing from or to your other core systems. That’s going to be a big challenge because, in a way, integration becomes an even bigger business constraint if you don’t have the right kinds of tools or approach to handle it.
The hype, of course, is around cloud and do you run on the cloud and can you integrate multiple clouds and all those kinds of things. We’re very heavily engaged and there is some very interesting innovation there, but the truth is most customers still have a lot of on-premise systems.
We’ve seen it around B2B as well, because there’s just more and more economic pressure to automate and streamline all kinds of B2B interactions that have traditionally been pretty inefficiently handled. Whether you look at it in the supply chain context where there is still a remarkable amount of sending of faxes between entities or you look at the health care arena, where health care providers really struggle with, not only patient records in one form or another, but also the whole complexity of multiple payers and multiple processors. And then internally, within their organizations, they're dealing with integration between lab systems and different specialists that may be running on different systems and so on. So all of those things contribute to the need for more integrations.
Lawson: We often talk about integration in this black box way - like it’s one problem you solve, when in fact, there are other related problems you have to tackle. Are there unforeseen, related issues organizations encounter when they get into the actual integration with a SaaS or even cloud computing?
Inbar: I think you touched on one major aspect of it, which is that all of us probably are guilty of thinking about it or presenting it as, “Here’s a black box, plug it in and you're done,” as opposed to owning up to the fact that it’s a process and it’s not going to stop unless your business stops developing.
I think in the SaaS world the thing that probably most customers underestimate - because nobody is really going to tell this to them – is every single SaaS application has to expose its data and processes through some kind of API, and as they do that, most of them are going to do that in a somewhat “unfriendly” or unpredictable fashion.
Salesforce happens to be one of the best that we’ve ever seen. Not only have they taken the development of their API and ability to access data and processes and everything else, they’ve taken it a lot further than anybody else in the market, but they’ve also put a lot of effort into trying to preserve consistency and backwards compatibility and so on. They're on version 10, I think, of their API. Every time they have a new release, we make sure that our Salesforce connector still works and takes advantage of new features and so on.
But it’s interesting how many of our customers keep running with the old connector, sometimes for years before they notice that, whoops, maybe they're missing something or something changed.
That changing API is a big issue if you're not as disciplined or as highly developed as somebody like Salesforce. With the newer vendors, a lot of them are still on the learning curve as regards this functionality. They deliver a version one, version two, version three of their API, but they can all be radically different because it’s hard to get the design right the first time. They may not get it right, exposing the data objects and business rules, for the next two or three times – and that can cause a lot of pain and slow down adoption and customer ROI.
I think that aspect of SaaS -- the maturity and the stability and the capability of the API -- is very much underestimated.
The other interesting area with SaaS – you know, security by and large is pretty good; availability by and large is pretty good. Handling larger data volumes over the Internet -- that’s still a moving target. There are some real constraints there that are steadily being tackled and there are various ways of approaching that.
Lawson: You mentioned security. Is the point of integration a security concern for people? Should it be? And are you, as an integration provider, responsible for that?
Inbar: You're generating more data silos in the sky and you're distributing data into many places as opposed to one place. So, those questions are coming up. Encryption is great, but in some scenarios you actually need to go beyond encryption and actually remove or add some fields or handle them in different ways to minimize potential data breaches or losses.
One of the other things we see, and we’ve done this with DataCloud, is that very often SaaS vendors want to be able to design integration effectively in the cloud, linking to their customer's systems. But when they deploy, while they want to monitor the integration and make sure the business rules are up to data and so on, you get back to this thing with integration agents where you don’t actually move the data off-premise. They actually manage from one cloud but move the data either within premises or to some separate, secure area that is viable. So this whole idea separating monitoring and updating and so on from the actual movement of data itself is another pattern that we think is going to become fairly common. And it will become common in other ways, where you may move data between secure clouds, but are still able to manage the integration separately from that.
Lawson: So what’s new for Pervasive this year?
Inbar: We’ve got some announcements that are going to be arriving in the coming months, but in terms of what’s been new this year so far we have what we call version two of the Data Cloud on Amazon.
This year we released Pervasive Data MatchMerge. Data matching is pretty much at the heart of just about every single data quality issue you see. There’s a lot of conversations that take place around, not just the regular name and address stuff, but identity resolution and what that means and how you make matches of people or products or events (when) you're dealing with a large number of attributes and you're dealing with degrees of confidence, not black and white. Our Data MatchMerge offering is built around handling handle large data sets or very complex matching.
In terms of other new things that we announced in the last few months, Pervasive Business XChange unit does EDI in supply chains. Instead of requiring participants to deliver their data in EDI format, which some of them can do that readily and some cannot and some can consume it well and some cannot - they're able to deliver end-to-end, going from SAP in one end into a corresponding application with their business partners and they're doing that by leveraging the connectivity we have. That’s a relatively new track that we anticipate is going to be a growing trend of the next few years.
