The open source community seems like a very likely place for a great deal of DevOps breakthroughs to rise out of, after all many of the products that live in that community address individuals who wear many hats: developers, system administrators, technology officers. Canonical, the company behind the popular Linux flavor Ubuntu, has been working to edge into the DevOps ecology for a while with tools that permit automation of operations via scripts and ease system administration with their IT automation tool Juju.
However, this is just the tip of Canonical’s iceberg. Former DevOpsANGLE editor Klint Finley published an article about Ubuntu’s current foray deeper into DevOps for the cloud with their newest release of Ubuntu Server 12.10:
Automation is particularly important for enabling elasticity — the ability to use varying amounts of computing resources as demand changes. Elasticity is one of the main ideas behind cloud computing — whether you’re looking at a public cloud like Amazon Web Services or in a private data center that mimics Amazon.
Originally JuJu only officially worked with Amazon EC2, but as of this week it works with any instance of OpenStack running on Ubuntu — including the HP’s public cloud service. Canonical also announced a new web-based user interface for JuJu this week.
With this move, Canonical is adding the capability not just the availability for OpenStack and HP’s public cloud to Juju. However, this is only the tip of the Canonical iceberg, as also reported, the company has also been moving into lower-power processors, such as ARM.
Much of this comes from a partnership between Canonical and the chip-maker Calxeda. Looking at this it brings to mind two things: the capability to produce lower-power, greener server farms that use less energy and cost less money to run; and also a goldmine for Big Data ventures that take advantage of Calxeda’s parallel processing capabilities on “microserver” archetectures.
Severs are expensive to build and maintain, and they’re very hot, keeping a server idle waiting to be used can expend a great deal of energy and money that any small outfit couldn’t hope to afford—making it necessary for them to lease cloud time—and even those enterprises that can run their own private clouds seek to use clouds to reduce overhead, not increase it. As a result, a turn to lower power ARM chips that are capable of running at lower power and coming to life faster from cold-shutdown could make the grade over having to idle a hotter chip constantly as it lay in wait to be used.
Even better, when it comes to ARM-based chips, not only can they run at lower power than their Intel brethren, they excel at hyperparallel processing necessary for Hadoop, and Big Data analysis, something that having elasticity and the capability to automate bringing more or less severs online to deal with a particular computation could come in very handy. We’ve seen the ARM architecture courted by graphics card makers like NVIDIA for this particular property at lower power; and looking at Calxeda’s recent foray into the server, the company is looking to aim at the Big Data market with ARM.
Cloud automation, lower power servers, and Big Data processing capabilities will give Canonical a giant boost in market share.
Right now, all eyes are on the Big Data market when it comes to business intelligence, even as it intersects with the cloud and software-as-a-service; big things are expected to come out of next weeks StrataConf where technologies that will convey data analysis, lower costs, and enable elastic processing will come to light and be discussed by industry experts.
Bringing Juju, the Ubuntu cloud, and the capability to run on ARM servers will definitely set Canonical in place to ride the Big Data wave as enterprise shops—both big and small—look to DevOps resources that will enable them to take advantage of this reduced overhead.