Decision Maker

Can You Trust Vendors in the Cloud Age?

From security concerns to whether a product (or vendor) will be around in three years, there's a lot to consider when picking your next tech vendor, esp. in the age of the cloud (and yes, you'll still need on-premise staffing).

• By Don Jones
• 08/03/2011

Google recently announced the discontinuation of Google Health, a service meant to be an online repository of individuals' health-care records. The general idea was that you retained online control over your records, making it easier to share them with caregivers when needed. This is the latest in a long wave of Google no-starts: Google Wave, radio advertising, Ride Finder, Answers, GOOG-411. It's a long list. It's not unsurprising, and I'm not specifically picking on Google: Companies discontinue products and services any time they're simply not attracting enough users.

Google is also famous for leaving products in "beta" for years at a stretch. While we may moan about Microsoft v1.0 products being "beta code" that we paid for, Google literally offers services for a long time that are clearly branded as "beta," a tag that used to mean "not ready for anyone to depend upon."

Again, I'm not trying to pick on Google specifically, because these practices have become widespread. These practices -- here-now-gone services, grossly extended "betas" and so on -- do raise some valid questions when it comes to business products: Whom can you trust? Who will be around in five or 10 years? Which services will last? Is it safe to buy into what a company is offering?

No Safety in Size
Even giant companies and their products can suddenly be pulled out from under your feet. Ask all the companies that spent time and money investing in, and building on, the Sun Identity Manager product -- a product that Oracle has now discontinued in favor of its own Oracle Identity Manager. Owning an "orphaned" product is never fun. Whatever new product you're offered almost always involves starting from scratch with a new deployment, new buildout, new training, new processes and so on. We're told we'll have to "migrate" to the new solution, which usually makes the process sound a lot easier than it really is.

Change is inevitable in our industry, of course, but you generally hope for a smoother path. Users of Microsoft Windows NT Domain Services were given actual migration tools to move to Active Directory, the successor technology. Exchange 2003 users can truly migrate to Exchange 2010 in a phased approach. SQL Server often offers in-place upgrades when new versions come out. We've learned to deal with that level of change; being told that a product or service is being completely discontinued and that we're more or less on our own -- well, that's rough.

Cloud Concerns
This becomes even more of a problem in the cloud-enabled age. If you make a bad selection of cloud providers, you could wind up having an emergency if they go out of business. In fact, some smart businesses are already considering "mitigation plans," not unlike their disaster recovery plans, designed to be put into effect if a cloud services provider goes belly-up. Such plans might involve a hurried download of data, a panicked search for a replacement provider, and a lot of downtime and pain in the middle. At least with orphaned products in your own datacenter, you can keep using them for a while; with an out-of-business cloud services provider, you're up the metaphorical creek without a paddle. Or even a breeze to push you along.

We investigate features and pricing, we conduct pilots, we argue about merits, and we eventually plunk down money as if we were buying an on-premises solution -- but we don't always appreciate that our new cloud toy could go away even more suddenly than it arrived.

Mitigation is the watchword of IT going forward. Decision makers who think that outsourcing anything in any way will eliminate on-premises staff are kidding themselves. Someone will still need to be in place to manage that solution, and you're going to need a plan in place to deal with potential disasters. With on-premises solutions, we usually understand the traditional disaster scenarios and can plan to mitigate them. But what would you do if a critical application were discontinued tomorrow? Does your company have mitigation plans in place to deal with suddenly orphaned products or cloud services? Can you afford not to?