The Riverbed Blog (testing)

A blog in search of a tagline

Cloud Balancing: Don’t Put All Your Apps in One Cloud

Posted by riverbedtest on September 29, 2011

In today’s post, we close our series introducing Zeus Technology and ADCs. Of course, now that Zeus is part of the Riverbed family, we’ll continue to write about ADCs regularly, especially now that we’ve given you a primer on the technology and what it does. Last week we talked about cloud bursting, so let’s close this series with another useful cloud trick that softADCs enable: cloud balancing.

Cloud balancing is the process of routing transactions and network requests across applications in multiple clouds. In plainer terms, it’s the simple “don’t put your eggs in one basket” approach – or in this case, don’t put all your applications in one cloud.

You might recall, earlier this year, Amazon Web Services was hit with a multi-day service outage on the East Coast after a “misaligned network” brought down several EC2 services in its Northern Virginia data center.  Or perhaps you remember when Microsoft’s Business Productivity Online Services experienced a multi-day email outage, with the culprit being “malformed email traffic on the service.” Cloud balancing acts as an insurance policy against such outages. For instance, Amazon’s EC2 customers could have avoided being impacted by the sweeping outage if they had implemented load balancing across two cloud providers, or with a cloud provider and their own data center. Yet too many organizations transition to the cloud without asking, "What if this provider suffers an outage?"

After all, even if we assume that all cloud providers can deliver 99 percent availability, deploying an application across several cloud instances can significantly decrease the chances that it will ever suffer from an outage due to the cloud provider.

Let me provide some perspective on this: 99% uptime means 1% downtime, which is equivalent to about 3 1/2 days a year.  Each new cloud provider adds another two “nines of availability” (raising the level to 99.99 percent, or about 52 minutes of downtime per year), so three platforms can deliver 99.9999 percent of uptime (six nines, or 31 seconds of downtime per year). From this point, the risk of cloud platform failure is effectively eliminated and the organization can concentrate on managing internal risks. Untitled

For another perspective on highly available system design, "the nines", and clustering (which is Cover effectively what this model emulates) may I recommend my book, Blueprints for High Availability, highly available at amazon.com and other purveyors of excellent technical books.

Because Zeus’ solutions are softADCs, they are designed to balance application traffic across multiple cloud deployments, reducing customers’ risk and improving the performance and capacity of applications. This allows customers to:

  • Increase the reliability of a cloud-based infrastructure by hedging the risk across multiple availability zones and cloud platforms.
  • Improve the performance of the cloud-based service using geographic traffic distribution and local traffic acceleration.

Many of the customers we’ve talked about here in this series, including Gilt Groupe and STA Travel, have deployed Zeus solutions to replicate application content across multiple public or private clouds and have seen measurable performance benefits as a result. Balancing application traffic across multiple clouds delivers confidence that the application will always be available, no matter what misfortune befalls any individual cloud provider. Because the reality is, you need some safety protocols in place with the cloud. In the wise words of ZDNet’s Ken Hess, “The cloud isn’t perfect, People. Computers make up the cloud. Computers are fallible. Therefore the cloud is fallible.” So, stay prepared and don’t put all your applications in one provider.

Leave a Reply

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

WordPress.com Logo

You are commenting using your WordPress.com 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: