Friday, February 25, 2011

AWS Cloud Formation

Amazon Web Services announced AWS CloudFormation, which lets developers and system admins use recipes to create and provision resources in Amazon cloud. This is conceptually similar to Opscode’s Chef Recipes which lets ops folks configure some aspect of the systems in their ecosystem. Clearly, AWS must have seen how Chef (and, of course, cfengine and puppet) has changed the configuration management landscape and wanted to do something which will help baking clouds much easier. CloudFormation is a right step in this direction.

With CloudFormation, developers can either use templates available in the library or create their own templates to describe AWS resources and the associated runtime needs of their application without worrying about the order in which the AWS resources should be provisioned and, also, making the provisioning work seamlessly with live applications without any disruption.

AWS CloudFormation supports many AWS resources including EC2, EBS Volumes, Load Balancers, Elastic IP, Security and Autoscaling Groups, Elastic Beanstalk, Cloudwatch Alarms, RDS, SimpleDB, SNS, etc.. They have also released recipes to install some of the open source applications like WordPress, Drupa, tracks, Redmine, Joomla, etc.. With just a few configuration details like the type of EC2 instance, autoscaling limits, etc. one can easily get these applications running in minutes.

AWS CloudFormation is definitely the next logical evolution for Amazon. This also gives them an opportunity to try and lock in their customers inside their ecosystem. If anyone expected them not to take this step, they are being naive about how business is done in this competitive world. Amazon is doing everything right to stay as the largest cloud player in the market. However, I do think that it doesn’t bode well for the players in the AWS ecosystem.

Some pundits see this as a direct threat to Chef and Puppet but it is definitely not the case. Chef and Puppet are more focussed on configuration management and are not reliant on AWS in any way. However, it does affect some large players like Rightscale and smaller ones like Bitnami . Even though their businesses are not entirely reliant on AWS alone, it does highlight the risk of any provider being reliant on a single cloud provider ecosystem, especially ambitious ones like Amazon. Do you think tt is yet another wake up call for anyone wanting to build a business around AWS Cloud?

Wednesday, February 23, 2011

Lets take Amazon recently announced a video streaming service available for free to their Amazon Prime subscribers. From what we have heard in the media, Amazon will end up offering a competitive service to Netflix. We all know that Netflix has recently moved most of their infrastructure to Amazon Web Services .

When Amazon announced the new video streaming service, cloud pundits raised some valid questions on whether Amazon will tweak the QoS of their cloud infrastructure to help their own service over Netflix. Clearly, in a highly competitive environment like what exists between Netflix and Amazon, this is a realistic possibility. Remember, in spite of the openness we tout in the cloud world, the cost of moving the infrastructure away from Amazon will be prohibitive for Netflix. Will Amazon play dirty games to kick a competitor away from the market or will it play straight and protect their booming cloud business.

As soon as this discussion came out among Clouderati (a loose group of cloud practitioners, vendors, pundits on Twitter), Amazon CTO, Werner Vogels, jumped in immediately and clarified  that Amazon has no such plans and he even highlighted that any such attempts will be shortsighted move and doesn’t bode well for the longevity of both Amazon Web Services and Amazon’s brand itself. He pointed out to how Amazon is already living in peace with competing third party merchants on their ecommerce platform. He even highlighted the fact that Amazon has a history of cannibalizing their own business to support a customer oriented view.

I agree with Werner’s arguments completely. Some of my thoughts on this development are:
Amazon is very smart to not cannibalize their larger brand for short term gains
Netflix is not dumb to consider such possibilities before they decided to put their entire business on Amazon infrastructure
More importantly, regulators are not blind. Any anti-competitive measure will not sit well with regulators and Amazon knows it pretty well

However, this does raise a very important question which every business planning to move their infrastructure to a third party cloud provider should consider. Are they willing to trust the cloud provider not to poach into their business? Do they have enough protection through their SLAs? Depending on the nature of the business, it is important that due diligence is paid while evaluating the risks.

Saturday, February 19, 2011

NoSQL is a fad

I attended a Silicon Valley Cloud Meetup this week where Siddharth Anand (Sid) gave a great talk on how Netflix has moved to the cloud from applications hosted in their own datacenter. In particular he focused on moving applications from using a traditional RDMS (Oracle in this case) to first SimpleDB (an Amazon offering) and then Cassandra (an open source Key/Value store). SriSatish Ambati (DataStax <http://www.datastax.com/> , has given a great overview of Cassandra’s history and capabilities (slides <http://slidesha.re/svcc_nflx> ) as a warm up talk.

Sid focused down on the actual issues their engineering team ran into when moving to the could, including:
  • Data Model changes required
  • Living without SQL support
  • No joins
  • No transactions
  • No triggers
  • etc.
Netflix accomplishes this by a common layer in their architecture that deals with the NoSQL and hides some of these idiosyncrasies from the application layers above. This is a lot of work (as evidenced by their recent hiring spree – they do these talks to recruit), but is necessary if they are to outsource their resource management and planning to the cloud so that they can scale efficiently.

Sid will give a followup part 2 talk that deals in more depth with the Cassandra issues that they are still working on solving. His slides <http://slidesha.re/hOkpT1> and a whitepaper <http://practicalcloudcomputing.com/post/1267489138/netflixcloudstorage> are available if you want more details.