Does Integration’s Heritage Matter in the Cloud?

IT Strives to Reinvent Itself Despite Budget Restrictions

A discussion is going on right now about the role of the enterprise service bus in cloud integration. Does it matter?

I’m not convinced it does. Most of the discussion seems to be coming from vendors, and while it’s probably good thought fodder for architects, I’m unconvinced there’s much of a strategic case for caring here.

One recent example, “Why Buses Don’t Fly in the Cloud: Thoughts on ESBs,” appeared on Wired Innovation Insights and was written by Maneesh Joshi, the senior director of Product Marketing at SnapLogic.

It’s a fine piece that recounts the successes and shortcomings of service-oriented architecture, and the ESB’s role in that.  I think it’s an excellent analysis of what’s happened in the past 10 years or so with ESBs and enterprise integration in general.

More to the point, though, is Joshi’s argument that cloud and SaaS adoption have “disrupted the legacy of software delivery model in a good way.”

Joshi’s thesis, in a nutshell, is that ESBs are the “old way,” based on older, more complex standards like XML. He’s arguing for a ‘next-generation’ approach that uses JavaScript Object Notification — more affectionately known as JSON.

Putting a JSON format on a traditional, XML-based ESB is like making a silk purse out of a sow’s ear, as he describes it. To that end, he recommends that you look closely at what’s going on with an iPaaS (integration platform as a solution) tool to see what’s really happening before you invest.

“Sticking with legacy technologies such as ESBs will only hamstring organizations from innovating rapidly and capitalizing on emerging opportunities,” he writes. “Companies must reconsider some of the old paradigms and embrace technological advances in integration to stay competitive and relevant in today’s fast moving markets.”

It’s a detailed discussion, but I walked away unconvinced, and here’s why: This is an old argument in the integration space that pre-dates cloud.

SnapLogic traces its roots back to ETL – extract, transform and load, which is actually even older than ESB technology. Meanwhile, competitors like MuleSoft and Dell Boomi rely on ESB. Both groups have evolved beyond their initial roots and offer solutions that integrate with the cloud, either cloud-to-cloud or on-premise-to-cloud.

In fact, Dell Boomi’s and Mulesoft’s solutions ranked as leaders in Gartner’s recent integration platform as a service (iPaaS) report. This would suggest to me that this isn’t a deal breaker for cloud integration.

I’m not making a technical call here. Obviously, you need to rely on your architects and systems people to judge which technology works best. It’s quite possible that nuances between these approaches matter deeply to your use cases. If you suspect so, then by all means, study what Joshi and others say about it.

Still, if I’m honest with you, to me, it sounds more like an old heritage debate that matters more to the vendors than to you.

I see this as another reminder that CIOs must focus on business needs and strategy more than the technology. By focusing on the business performance requirements and needs, you can worry less about which technology does what and more about solutions that work for you.