This week in cloud: hedging Amazon’s regions and Rackspace seeks deals

#SuryaRay #Surya Last week, RedisToGo started making its hosted version of the Redis key value store available from Amazon’s US-West facility data centers to all its customers. This may not seem like ground-breaking news, but it’s worth noting since RedisToGo was reliant on Amazon US-East, the giant cloud provider’s oldest and biggest data center complex which has been subject to several glitches over the past year. Those issues have have taken down websites running on AWS including Reddit, Foursquare, and Heroku and led many to wonder why these tech-savvy companies leave so much of their overall workload in that problematic facility.

According to the RedisToGo blog:

“2012 has been a rocky year for AWS-East. We’ve seen several outages from freak storms to gremlins in the servers, making it hard for anyone wanting to achieve 99.999% uptime. We’ve been planning multi-zones for a while, but as Hurricane Sandy approached AWS-East, we were ready for the worst. We sent a mail to our customers and opened up our multi-zone to all customers as a manual process.”

After Superstorm Sandy — which did not impact US-East after all but wreaked havoc on other data centers in New York and New Jersey — RedisToGo offered the US-West zone to all its customers.

As reported before, there are reasons that many companies relied so heavily on US-East. Some Amazon services debut there first and cost less than in other regions where higher labor and utility costs are passed along to customers. But, given the downtimes suffered by these AWS customers, you have to wonder if all of them are considering a similar move or even a move of some of their workloads to a second cloud provider.

Rackspace goal: Converting tire kickers to buyers

Rackspace, which would very much like to displace Amazon in large accounts, is happy with the interest it’s gotten so far in its 3-month-old OpenStack-based cloud compute offering. Now it needs to convert that interest into actual deals, Rackspace CEO Lanham Napier told me in an interview after the company’s q3 earnings call on Tuesday night.

San Antonio-based Rackspace is the godfather of the OpenStack open-source cloud which is now managed by the OpenStack Foundation and in the past few months it’s rolled out more pieces of that private- and public-cloud puzzle.

On the public cloud side, Rackspace will take on Amazon, at least in service conscious shops that want and need support and service. But it will also face off against a growing number of its OpenStack brethren — Hewlett-Packard, Cloudscaling, Internap, Red Hat and others. But the field is even more crowded than that with more established open-source cloud technologies from Eucalyptus, CloudStack and OpenNebula also in the mix.

Centurylink – Savvis goes big on IaaS

When Centurylink bought Savvis two years ago, the game plan was to combine Centurylink’s telecom expertise with Savvis’s data center know-how to provide cloud infrastructure — which it’s been doing for enterprise customers in a custom one-off manner. Now, the company’s SavvisDirect service, in beta since last month, provides on-demand, self-provisioning infrastructure by the hour, according to _InformationWeek_.

While AWS remains the dominant player in public cloud infrastructure, it’s still early yet in the world of cloud migration. The latest Gartner Magic Quadrant ratings shows that dominance — Gartner gives Amazon its due:

“AWS has by far the largest pool of capacity, which makes it one of the few infrastructures suitable for batch computing, especially those workloads that require short-term provisioning of hundreds of servers at a time. AWS also offers specialized infrastructure options for high-performance computing and big data applications, along with a “spot pricing” market for compute capacity.”

But AWS is not invulnerable. Critics, including Gartner, say its narrowly-defined Service Level Agreements (SLAs) leave much to be desired by enterprise customers and that the bewildering array of separately priced cloud services is confusing.

Probably most important, there is an array of savvy IaaS competitors coming down the pike including Terremark,  CSC, Joyent, Tier3 Networks, Virtustream and others, Gartner said. @suryaray

RedisToGo interview - Simple Redis Hosting

We’re back! After a week off, we’ve interviewed Ben Arent of RedisToGo.

1, Please, tell us about your project

Simple Redis Hosting

2, Why did you choose to work on this idea?

We loved Redis and couldn’t find a solution that worked for us.

3, How many hours do you work on your startup per week?

Full time.

4, What works for you in terms of getting users?

We are a featured add-on on Heroku, but are getting increasing amount of users from our blog

5, Looking back 6 months, what would you’ve done differently?

Explored more creative product ideas. We’ve found our Redis Enterprise a bit hit.

6, What’s the hottest startup these days, and why? (your personal opinion) - Cause Cats.

Thanks again to Ben from RedisToGo for finding the time in his busy schedule to do this interview!

June 30th Downtime Postmortem
At about 6am Swiss time we suffered an outage because of multiple issues that formed a chain reaction.
It all started with severe storms in the North Virginia region of the US. Because of the weather conditions the Amazon Datacenter located there had power failures. Which in effect led to Heroku and RedisToGo, two providers sharelephant is relying on, having issues themselves.
While Heroku, the service running our application, managed to recover quite fast with only sporadic downtime windows of 15 Minutes this was not the case for RedisToGo our Session storage. The sharelephant server could therefore not connect to Redis when starting the application and was failing for multiple hours.
We are already working on a plan in order to avoid such outages in the future.
Sorry for any inconvenience you had to endure.