Like Stephen Hawking and his search for the elusive Theory of Everything, data-center technologists have a quest — code that will run “everything” in a data center. And if recent announcements are as they seem, data-center operators may have their wish.

Some well-heeled venture-capital groups — Khosla Ventures, Andreessen Horowitz, Fuel Capital, and SV Angel — appear to think so having provided Mesosphere Inc., the company behind the data center “everything code,” additional venture funds amounting to $36 million. Besides announcing the latest round of funding, the company introduced the Mesosphere Data Center Operating System (DCOS). This is “A new kind of operating system that spans all the servers in a physical or cloud-based data center, and runs on top of any Linux distribution.”

The backstory

In an earlier post, I made mention of paying attention to what Amazon’s data-center guru James Hamilton says. That same scrutiny applies to Google Fellows Urs Hölzle and Luiz André Barroso — two reasons why Google data centers are cutting edge, and why frugal me spent $26 on their 2009 book The Datacenter as a Computer: An Introduction to the Design of Warehouse-Scale Machines. The missive introduces the concept of the data center as a computer. “The computing platform of interest no longer resembles a pizza box or a refrigerator but a warehouse full of computers,” wrote the authors. “We must treat the data center itself as one massive warehouse-scale computer.”

It’s not much of a reach to assume Google (hint: BORG), and the other big-name data-center operators have their own DCOS, but they’re keeping quiet, which leaves other and smaller data-center operators to fend for themselves — until now. Enter Mesosphere, a DCOS available to everyone (using Linux servers), that runs in both private data centers and in public clouds.

What is a DCOS?

Data-center management systems, which is what most data centers now use, focus on controlling physical systems. DCOSs are the next evolutionary step, in that DCOSs can control a data center’s logical and physical systems. The University of California, Berkeley research paper The Datacenter Needs an Operating System (PDF) describes what a DCOS brings to the table.

  • Resource sharing: DCOS can multiplex resources between users of an application and across applications.
  • Data sharing: Besides resources, DCOSs traffic data between the necessary applications.
  • Programming abstractions: DCOSs provide user interfaces that hide the intricacies of hardware and simplify application development.
  • Debugging: Figuring out what massively parallel applications are doing remains one of the hardest challenges in cluster computing: DCOSs use correctness and performance debugging to address the challenges.

Mesosphere DCOS features

The Mesosphere DCOS is built on top of Apache Mesos. The website mentions, “Mesos installs with two different components, the master and the slave. The master is a process which runs on a node in the cluster and orchestrates the running of tasks on slaves by receiving resource offers from slaves and offering those resources to registered frameworks.”

That allows the DCOS to run the entire data center as a single entity, allowing administrators to engage servers and software as they need or configure either servers or software to activate automatically as requirements change. This YouTube video offers a look at administrating a data center using the Mesosphere DCOS.

Some features built into the Mesosphere DCOS are:

The Mesosphere website mentions, “The Mesosphere DCOS will be available beginning 2015. If you want early access, sign up here.”