This week, Amazon launched a feature called OpsWorks that handles the application life cycle management deployed on the cloud. This fills the critical gap that exists between initial migration of the applications and managing the on-going operations and change management of cloud applications. Irrespective of the size of deployments, OpsWorks makes the change management extremely easy.
According to Dr. Werner Vogel’s latest blog post:
“OpsWorks allows you to manage the complete application lifecycle, including resource provisioning, configuration management, application deployment, software updates, monitoring, and access control.”
But aren’t the attributes that Werner mentioned are the key features of PaaS? Platforms like Heroku and Engine Yard (Both run on AWS) promise the exact same set of capabilities to developers. As a deployment environment built on top of Amazon building blocks, even AWS Beanstalk delivers the same abilities. Though Amazon has positioned OpsWorks as a different tool from BeanStalk, the end goal is the same: to reduce the complexity of dealing with the cloud application deployments.
AWS OpsWorks announcement surfaces a number of interesting facts:
- Though not broadly published, OpsWorks is a reincarnation of Scalarium, a product developed by Peritor. AWS acquired Peritor in 2012.
- Amazon wants to create a sticky environment that makes it hard for the customers to get out of its fiefdom. Once a multi-tier application gets deployed through OpsWorks, it is complex to move it out of AWS. Customers will not be dealing with plain vanilla EC2 VMs anymore, and with OpsWorks they get entrenched in the AWS world.
- With OpsWorks, OpsCode Chef officially gets the blessings of AWS. Though it’s been possible to integrate Chef recipes with CloudFormation templates, this is the first time that AWS talked about Chef as the preferred automation tool, leaving aside its potential competitor, Puppet (see disclosure).
- It’s also interesting to note that the preferred monitoring tool of OpsWorks is Ganglia, not Nagios. This is again the first time that Amazon has endorsed a third-party monitoring tool alongside CloudWatch.
- Instances deployed through OpsWorks need to be at least m1.small because of the resource requirements of Chef. This means no support for t1.micro, which will force customers enjoying the free tier to pay.
The announcement for OpsWorks has a significant impact on the cloud ecosystem, and sends a strong message to the AWS partners that Amazon is keen on going up in the stack and also broad with its service offerings.
OpsWorks will affect the following players:
- PaaS players. Agree or not, OpsWorks will compete with the potential PaaS players like Heroku, Engine Yard and AppFog. The flexibility, deployment experience and the time taken to go live positions it as a viable alternative to the popular PaaS offerings in the market. So, Salesforce and Heroku should really take this new product seriously.
- Cloud management tool vendors. The first impression that I got when I read about OpsWorks is that it is Scalr. The terminology, experience and the workflow are very similar to Scalr. In Scalr, it starts with a Farm and then individual Roles are added which are launched as Servers at a later point onto which the Applications can be deployed. AWS OpsWorks follows almost the same nomenclature of defining the Stacks and configuring individual Layers and adding Instances on which the Apps get deployed. Even the Chef integration and custom recipes look similar to Scalr. Though most of the management tools target multiple cloud platforms, this move from Amazon will hurt them in the long term. Going forward, Amazon might even add a graphical interface supporting the drag and drop of layers to compose an application deployment. Transcend Computing, are you listening?
- Cloud consulting firms and medium-sized system integrators. Many customers are not comfortable configuring ELB and dealing with the EC2 auto scale parameters. They approach niche cloud migration companies who happen to be listed on AWS Partner Network. With OpsWorks, even the entry level Sys Admins will be able to handle complex deployments. While this is great for customers, it will take away the migration and operations business of the partners.
This is not the first time that Amazon has intimidated the ecosystem. As my colleague from GigaOM, Barb Darrow pointed out a year ago, with the launch of each new service, Amazon distances itself from a set of partners. The recent Redshift launch has sent strong signals to Oracle, IBM, and Terdata. When AWS added features like cross-region EBS snapshots and Trusted Advisor, the company disrupted Ylastic’s business model. Amazon is doing exactly what Microsoft did in its heyday: making a set of partners look redundant with every new release.
Disclosure: Puppet Labs is backed by True Ventures, a venture capital firm that is an investor in the parent company of this site, Giga Omni Media. Om Malik, founder of Giga Omni Media, is also a venture partner at True.