Inside view of Tibco integration architecture planning

They may be called technology boutiques, though some prefer the term "best-of-breed" tool providers. Whatever they are called, a parade of innovative software startups have become part of Tibco Software's portfolio.

A long-standing player in rapid-fire enterprise middleware, Tibco began a move into business analytics in earnest in 2007 by buying Spotfire. Tibco followed that up with a series of acquisitions of such vendors as StreamBase Systems, Jaspersoft, Mashery, Statistica, the former Composite Software data virtualization business of Cisco and others. The vendor's buying spree was a little diminished when, in 2014, Tibco went private and left the ranks of public companies.

For Tibco, last year's acquisitions of Scribe Software and Orchestra Networks, as well as this March's purchase of SnappyData, are part of a drive toward what it calls connected intelligence -- that is, enterprise middleware that enables operational analytics at an ever-larger scale.

We spoke recently with Nelson Petracek, CTO at Tibco, to find out how the company is riding the waves of cloud, messaging and event processing innovations on the way to next-generation integration architecture.

We've lately seen data velocity, variety and volume gain new influence in integration architecture. How does that play out as Tibco plans technology strategy?

Nelson Petracek

Nelson Petracek: When you look at some of the changes that have occurred since the days of what people called enterprise application integration, you've seen shifts. The shifts particularly are in how organizations approach where and how they want to deploy software. People are moving into hybrid cloud environments or even multi-cloud environments.

A lot of the conversation is around the speed with which you deploy new capabilities -- how you build reusable components, using APIs and API contracts to simplify how you actually integrate with different applications and systems -- [and] how you expose those to the outside world.

What was behind the thinking of Tibco's recent purchase of SnappyData?

Petracek: Well, we find that people are typically using [Apache] Spark to handle their ETL-like [extract, transform and load] workloads. But one of the challenges that we saw was in people's ability to achieve speed to access that information. With SnappyData, you represent information in such a way that you access it more quickly. It allows you to access the information in a Spark cluster -- with big data that people have collected over the years -- more efficiently.

Could you provide some context for last year's acquisition of Scribe Software? What does that bring to users?

Petracek: Scribe is really an acquisition meant to enrich capabilities in the interconnect area. It's targeted at a specific, or more focused, persona. It's not for the person that's going to be writing a Linux kernel over the weekend; instead, it is more of a model-driven approach for a different type of user.

We cover developers, but, on the other side, we wanted to focus more on not necessarily technologists, but people that still had integration requirements, people who just wanted to wire [applications] together. That's where Scribe is focused -- not on a hardcore developer, but someone who still needs to move and integrate data and systems together.

Let's look at Tibco's Orchestra Networks acquisition. Is that a move deeper into the data side of integration architecture?

Scribe is really an acquisition meant to enrich capabilities in the interconnect area ... It's not for the person that's going to be writing a Linux kernel over the weekend.

Petracek: Yes, Orchestra has traditional master data management capabilities, but we actually see it as a much broader capability than that -- one that is useful for organizations that are realizing that they're going to have a bigger problem when it comes to metadata and reference data management.

People now worry about what information they actually collected in their big data environment -- how old it is, how fresh is, how accurate is it and where it came from. A lot of people just don't have a handle on it. They collected a bunch of data, but they don't really necessarily know what's in there, where it came from [or] how good it is.

Would you say the changes in integration architecture resemble changes going on in databases, where fit-for-purpose NoSQL databases have become common?

Petracek: Yes, we view messaging as similarly taking almost a polyglot approach, just like you see with databases. Yes, you're going to have a relational databases for certain use cases, but you will also have key-value stores, and now we're seeing a lot of renewed interest in graph databases.

Just like you're seeing all these different databases, we're seeing messaging come to include microservices capabilities and even event processing capabilities, because now people have more event-oriented information. They will need to handle streams of data that come from any source at any time.

What approach do you take to the new customers you gain, ones that weren't joint customers of yours and the companies you've acquired? How do you nurture the software you acquire?

Petracek: It's a matter of consideration for customers whenever any technologies are acquired by another organization. So, we do spend a lot of time making sure that people know who Tibco is, that we have been around a certain number of years and have a number of key customers across the globe. We need to make sure they understand what our roadmap is.

We've got our homegrown innovations, and we've got products that we've acquired, and we have an innovation lab that both use. And the lab can focus even on technology that might be further out -- things like blockchain, natural language processing and IoT.