TDWI Upside - Where Data Means Business

A New, Converged Approach to Data and Application Integration

Think of "hybrid integration" as a new paradigm for data and application integration. It's tailor-made for enterprises in which IT resources live both on premises and in the cloud.

Think of "hybrid integration" as a new paradigm for data and application integration. It seems tailor-made for enterprises in which IT resources live in both on-premises environments and the cloud.

This could make it just the thing for the enterprises of the present and foreseeable future.

"Hybrid integration" is recognized as a category by market watchers such as Forrester Research and Gartner. The latter published its assessment of the hybrid integration space earlier this year.

In November, Forrester weighed in with its latest take on the hybrid integration space.

The Forrester Wave: Hybrid Integration for Enterprises, Q4 2016 breaks down the market for hybrid integration products and services. Forrester sees a dozen or so vendors -- led by Informatica, Software AG, Dell Boomi, MuleSoft, and IBM -- jockeying for space in a rapidly maturing market.

It's a fast-growing market, too, according to Forrester's Henry Peyret.

"The hybrid integration market is growing because more [enterprise architects] see the broad capabilities and use-case flexibility as a way to address key business challenges. This market growth is in large part due to firms rethinking their integration strategies," writes Peyret.

Hybrid Integration Explained

Hybrid integration uses technologies and concepts specific to both data integration (DI) and enterprise application integration (EAI). These include ETL and data federation (mainstays of DI) and event processing, the enterprise service bus, and services-oriented architecture (mainstays of EAI).

According to Peyret and Forrester, hybrid integration is more or less what it sounds like: a means for knitting together the on- and off-premises resources of the distributed enterprise. The on-premises pillars of most companies' data and application integration strategies -- DI, EAI, ESB, SOA -- weren't designed for hybrid integration scenarios, he argues.

"[T]he integration landscape has changed from solely on premises, with integration points being slow-changing, to a much more varied landscape that includes cloud for systems of engagement and systems of insight -- as well as the expectation that as these [systems] rapidly change, the integration interfaces will also rapidly change," Peyret writes.

The upshot, Peyret says, is that enterprise architects are looking for alternatives to EAI, ESB, SOA, and other on-premises-oriented technology paradigms. As Peyret sees it, they have two options: in the first case, they can complement their existing EAI, ESB, and SOA tools with integration-platform-as-a-service (iPaaS) technologies. At the very least, this will give them connectivity to the cloud. This isn't what Forrester means by "hybrid integration," however.

Enter option two, an approach in which companies invest in new hybrid integration platforms. The idea in this case isn't so much to replace existing ESB and SOA investments but to "renew" them -- i.e., by transforming them into "[f]ederated on-premises and cloud-based integration platforms."

Retrofitting for Hybrid Integration

This is easier said than done. As Peyret and Forrester note, even though the available vendor offerings are maturing rapidly, they still aren't quite ready for prime time.

Hybrid integration offerings are lacking in five critical areas, according to Peyret:

  • Application and data integration are converging, but the available offerings don't yet address this.

  • Their support for critical management functions -- debugging, alerting, troubleshooting, etc. -- is sorely lacking. Peyret says the available functions are the "weakest aspect" of current offerings.

  • Few if any vendors offer the equivalent of a "complete" cloud architecture. Peyret says such an architecture should be completely Web-based for critical functions (development, management, runtime, etc.) and designed for microservices. It should offer support for container-based deployments and share a single code base between the public and private cloud, as well as the on-premises enterprise.

  • Few offerings support the self-service data prep use case. Peyret calls this the "citizen integrator" use case. "We see an increasing need to support technically savvy business users ... as a way for technology teams to avoid being a bottleneck for simple integrations. Most hybrid integration solutions remain complex and require integration specialists," he writes.

  • Security is made even more complicated because different verticals have different security requirements.

Hybrid Integration Tools Still Improving

In the end, Peyret indicates, hybrid integration platforms, tools, and services are getting better.

For example, every vendor Forrester surveyed says it "is investing to continue to bring a better support of the cloud architecture on one point or another." Elsewhere, the use of machine learning and artificial intelligence technologies could help ease the (significant) shortcomings of existing management functions.

Lastly, enterprises are pushing vendors to address the convergence of EAI and DI. At least one data integration player has taken up the challenge, Peyret says. "[E]nterprises that have already invested in data integration are expanding their footprint to include application integration -- vendors such as Informatica leverage this shift to expand their markets."

TDWI Membership

Accelerate Your Projects,
and Your Career

TDWI Members have access to exclusive research reports, publications, communities and training.

Individual, Student, & Team memberships available.