It sounds incongruous. Chatter is associated with the Force.com platform, which runs in Salesforce.com data centers, not Amazon's. But 99% of CloudAware's customers are using Amazon cloud services, and it's built monitoring and management systems on Force.com to track their workloads there. Now, if there's something going wrong with your workload, chances are one of your Chatter feeds will contain the information you need to take care of it.
If you're not aware of CloudAware, don't feel bad. The 10-employee, Manhattan, N.Y., engineering consulting firm has been helping companies migrate workloads to the cloud since Amazon Web Services first went public in 2007. It saw the same tasks recurring so many times it decided to produce a service that captured best practices and automated execution of those tasks.
"We've been operating in stealth mode. Most people are not aware of us. We didn't have a website until six months ago," said Mikhail Malamud, founder and CEO of CloudAware.
Malamud expects CloudAware to eventually support Google Compute Engine and Microsoft Azure, but said so far "we haven't seen the demand for them." Also, from the start, the firm was focused on the early activity on AWS's EC2 and related services and built capabilities for monitoring and managing them. The CloudAware service, for example, implements about 50 best practices derived from Amazon's own Trusted Advisor, the rules engine that helps customers configure and deploy workloads. It's added 50 best practices policies of its own to help customers get their applications up and running with the least difficulty.
[ Want to learn more about Amazon's own Trusted Advisor? See Amazon Tool Helps Shape Your Cloud Workload. ]
CloudAware didn't create the service entirely on its own. Two Fortune 500 companies financed CloudAware's R&D to create the system so that they could use it. That means CloudAware, which hasn't yet taken a penny of venture capital, has a launched a service that counts Thomson Reuters, Reed Elsevier, Pearson, Sony, AOL and others as paying customers, and is breaking even. Revenue has increased 150% at the privately held company since the start of the year, although Malamud won't disclose what that amount is.
The service charges customers 2-7% of their total Amazon bill, depending on the number of service components that they use. It also charges a minimum of $1,500 a month. There's no freemium version with CloudAware. "We're going after the enterprise," not small businesses, startups or mom & pop e-commerce sites, he said.
The service includes the open source configuration engine, Puppet, along with other deployment components, so customers may use it to actually deploy their workloads as well as monitor them.
The use of Chatter kicks in as multiple users at an enterprise start deploying workloads to Amazon through CloudAware. It ties a set of security policies to a security feed in Chatter, and an incident with any workload triggers a message on Chatter to those receiving the feed.
Likewise, feeds can be set up for compliance, intrusion detection, governance, backup, cost control and other areas of workload management. If a developer changes a production system, that event would trigger a message on the Chatter governance feed. The systems administrator could then ask the developer through Chatter (perhaps from his mobile device) why he had made the change.
"Everything is out in the open. You can no longer sneak a change in. Everyone knows you did it, and you'll get scolded if it doesn't work. It generates a culture of responsibility," said Malamud.
The Chatter feeds can cross boundaries and siloed skill sets that still keep one side of the data center from talking to the other. Instead of just generating talk, the service also allows customers to set up approval processes so that a change to a running system goes through a set process. It offers 50 prototype processes that a customer may choose from to set up its own process.
When a virtual machine is generated or a workload is launched from CloudAware's app store, the service cross references the statistics that stream from Amazon's CloudWatch monitoring service with the information it collects and stores in its own repository. Analytics are applied to the operational data with the goal of spotting trouble as it develops and getting a "case" launched via Chatter to alert the responsible parties.
CloudAware tries to go farther than Amazon CloudWatch by monitoring information on the running application. "Amazon doesn't understand the concept of an application, whether it's in production or development," said Malamud, but CloudAware does. It includes tagging and labeling that lets a system administrator to develop a more detailed view of the running workload than can be obtained under CloudWatch.