There’s More Than One Cloud in the Sky!

Due to recent, and highly publicized, outages, there has been much commentary about the “validity” of the cloud. On the flip side, there’s been some great commentary on how such outages can be minimized significantly by leveraging the cloud.

I’ve always said that cloud services will never 100 percent replace an existing enterprise environment. There are many reasons why this is the case, but one I think is often overlooked is the amount of redundancy typically seen in today’s enterprise. It’s rare for a customer to have a single location where the company’s infrastructure resides. Why then do we seem so uncomfortable leveraging more than one cloud provider for consumption of cloud services?

Of course I’m biased when it comes to the company customers should choose as their cloud provider of choice, but I also believe there is enough reason and ample justification to explore how companies could leverage multiple cloud providers when delivering an application.

In the AT&T Proof of Concept Lab, we’re showing demos of how to load balance between our managed cloud services of Synaptic Hosting and our traditional customer managed cloud services, Synaptic Compute-as-a-Service. This concept could play out with services delivered by AT&T on one end and any number of other cloud providers on the other end. It all comes down to simplicity or complexity: How much time/money do you want to invest to ensure the overall reliability of your environment? A single cloud solution is much easier to architect and maintain than a multiple cloud solution, but is it worth it?

In a recent post, “Building an Intelligent Cloud,” our CTO, John Donovan, discusses AT&T’s cloud services and what makes us different. John’s points—such as the global interoperability of AT&T’s cloud services as well as the fact that they are localized and distributed—are spot on. In another blog, the CEO of a photo-sharing site explained why the site experienced only minimal effects due to the outage of its cloud provider, including that the site’s services are spread across multiple availability zones and that the company “designed for failure from day one.”

No service will be available 100 percent of the time. You have to plan for downtime. You can build a solution that either accepts there may be some level of downtime that “just happens” due to infrastructure or user error, or build a solution that can withstand these types of unplanned issues. Both are possible – it’s really up to you to decide what’s most important when thinking about which cloud you use to deliver your applications.

Tell me what you think!

What’s your cloud strategy?
Are you leveraging more than one cloud provider today?
Are you reconsidering whether or not this should be a requirement going forward?
The Networking Exchange Blog Team About NEB Team