The Riverbed Blog (testing)

A blog in search of a tagline

Cloud computing for government, part 2 of 3

Posted by riverbedtest on September 26, 2011

Cloud-computing-panel Today we resume with part two of my three-part series on government cloud computing. Be sure to read part one, in case you missed it.

The benefits of the cloud are supposedly self-evident, but how can agencies actually measure the ROI?

Curiously, in General Alexander’s testimony [see the last question in part 1], while he praised the capabilities of cloud security, he questioned some of the promised economic benefits. Many providers publish online calculators that allow you to compare the costs of a cloud deployment to the costs of running on-premise infrastructures. Frequently these fail to account for the personnel costs of installing and maintaining on-premise equipment, so in one sense they aren’t as good as they could be. However, trying to measure cloud ROI and comparing that to traditional infrastructure ROI ignores the cloud’s most important benefit: elasticity. The cloud allows you to add and remove resources according to demand. Traditional on-premise infrastructures are either under-utilized (and thus waste resources) or over-subscribed (and thus perform poorly). Applications designed to take advantage of the cloud’s elasticity largely eliminate the guesswork associated with predicting demand. A resource availability curve that always matches your demand curve appears a lot like perfect ROI.

What does Riverbed bring to this space that sets you apart from others?

Over eight years Riverbed has built a reputation of making wide area networks feel like they’re right next door. As organizations consolidated dispersed branch office resources into fewer large data centers, our technology has helped eliminate the typical problems that arise from computing at a distance. The cloud is a natural next step for us, because in many ways the cloud is similar to a WAN. Users can be situated anywhere and we can apply the same optimization techniques to make applications feel local. With Steelheads of various flavors you can vastly accelerate the movement of data from on-premise to the cloud and back, and also between clouds—even if the providers are different. In our Whitewater appliance we’ve adapted our optimization technology to remove the drudgery from backups, allowing you to point your backup software to a target that in turn compresses, encrypts, and backs up to the cloud—no more tape. For cases where you aren’t able to deploy our flagship symmetric optimization technology, we offer application acceleration that you can add to your cloud-based applications through two recent acquisitions: Zeus and Apptimize. And soon, through our partnership with Akamai, you can accelerate third-party SaaS applications by optimizing the high-latency link between your current location and a point of presence topologically very close to your ultimate destination. Regardless of which cloud providers you choose and what technology they’re built on, we can make any cloud perform better.

Is the cloud necessarily a permanent solution? When does it make sense to use the cloud as a temporary resource?

This would seem to conflict with the “cloud first” mandate and the notion that the cloud is the new default. It can be tempting to consider the cloud as an extension of an existing data center. Unfortunately, such thinking imposes limits—you’re less free to build applications that incorporate full cloud functionality and you can’t move to a full scale-up/scale-down resource curve. Also, I think this can create a mindset where the cloud becomes that “extra” thing that ends up not being managed well, or at all.

Is moving to the cloud strictly an IT issue? What other stakeholders need to be included in the discussions, and why?

IT organizations that choose, on their own, to move production workloads to the cloud do so at their peril. Capacity planning and disaster recovery require input from the agency’s working units. Data location and portability require consultation with legal and compliance teams. Cloud provider procedures and certifications require review by internal audit groups. Budgetary changes require working with finance folks. Don’t allow cloud projects to become line items in some developer’s monthly expense report!

Agencies will need to develop applications and services for their specific needs. Does the cloud change how they do that?

There are fundamental differences in the way applications should be built to run on clouds. Probably one of the most shocking changes is that servers are now disposable horsepower. Infrastructure is code: when you need compute resources, you simply issue a few API calls and within a matter of minutes those resources are available. Vast amounts of distributed storage are also there, waiting for you to allocate and use. In many cases this storage incorporates automatic replication, so you no longer need to build that into your application. Cloud computing also simplifies the process of updating applications: you clone an existing application, add and test updates, then move users over to the new version. Cloud providers often publish detailed technical guidance for how to develop on their particular platforms.

Part three will follow two weeks from today.

One Response to “Cloud computing for government, part 2 of 3”

  1. Dl380g6 said

    Usually I do not post on blogs, but I wish to say that this post really forced me to do so! Thanks, incredibly nice article.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: