We received multiple calls and emails about recent M&A in cloud integration and the hot issue of the right integration model in a cloud computing world. Also, Kim Nash had a great round-up over at CIO magazine.
First, a bit of history. Â In a former life, I co-founded Data Integration company Informatica in 1992 in the proverbial garage (too drafty; we soon moved it to a spare room) on the assumption that client-server computing would prevail over the mainframe and indeed we were right about that in spades. However we were dead wrong about one thing.
I was dead wrong to assume that the mainframe was going to die in the last century. I was in my twenties, Windows 3.0 was out, downsizing to Open Systems was all the rage and now ahem, certainly older and perhaps wiser, I’m here to tell you that the mainframe will outlive me and all the Unix revolutionaries who predicted its demise.
SnapLogic was founded in 2006 on the premise that cloud computing shall prevail in this decade. The learning of the past decades is that integration with on-premise systems will be required by businesses large and small. Â Whether itâ??s a small business using QuickBooks, or a large business using a legacy system with sensitive financial data on premise, these systems are not going away anytime soon.
Likewise integration has to mirror this hybrid computing reality and be available on premises, in a public cloud for many or in a private cloud for Fortune 50, tech savvy customers. Â As long as the integration technology is web-savvy and thus able to deal with versioning intelligently, and scalable enough — having your head in the cloud but feet on the ground may be the best policy.Â Witness how iTunes does this very well in the consumer world where an on-premise app links to a big multi-user store in the cloud for over a Billion users.
As long as you have on-premise apps you’ll need an integration element on premise (often called an agent, or atom) and then you’re “slightly pregnant” with versioning, and whether your integration solution is multi-tenant or not youÂ have to grok app versions and make it a core part of your value proposition to be successful in integration in this decade.
Certainly multi-tenancy helps keep the cost of integration down for integration vendors and when those savings are passed on to a customer its a good thing and we support that, but to suggest that muti-tenancy is only way to fly is to make the mistake of assuming that the mainframe will die!
Architecture matters. A strong, simple product that lets you easily tackle your integration matters. An open published API matters. That API used by the vendor themselves, creates an open integration platform which makes for cloud and cloud-to-ground integration solutions that fit well in any IT mix. Â This lets you choose where to develop, test and deploy your solutions whether on premise, in a public or private cloud, based upon the best fit — and value — for your company.