Now listen, Gartner – virtualisation and containers ARE different

Magic Quadrant starts to lose its sparkle

Comment Gartner recently released its Magic Quadrant for x86 Server Virtualisation Infrastructure. In it, the mega analyst lumps together hypervisor-based virtualisation and containers. This is wrong, and as I've discussed before virtualisation and containerisation are different.

Even if you consider all the differentiators in the hyperlink above to be meaningless technobabble, there are solid business reasons to separate these out. The really short version is that hypervisors are the past and containers are the future.

Hypervisors are a means of giving redundancy, mobility and capabilities to legacy 'monolithic' applications that those application were never originally designed to have.

We've been writing applications for decades that simply presume they are the only thing on a given system and that the system they are on will never fail. Those applications are what hypervisors are here to support.

Modern applications are designed to be "microservices". They are smaller, with each component broken out into its own separate mini-application.

The whole thing is designed to fail (in whole and in part) and recover from said failures gracefully. Containers are the technology that make these modern applications work.

You can put some legacy applications in containers. You could absolutely put microservices onto their own operating systems and shove them into hypervisors. Both things are completely absurd practices.

Attempting to conflate containers and hypervisors is tantamount to lumping together New Horizons' kilobit per second interplanetary radio system with 40 gigabit Ethernet under the guise that they are both "communications systems". No. Just no.

Who do you love?

In a sane world where containers and hypervisors were separate this would remove Odin (formerly the Parallels' Service Provider business) from the Gartner x86 Server Virtualisation Infrastructure Magic Quadrant. Now don't get me wrong, Odin is amazing.

I've taken the time to investigate the Virtuozzo containerisation offering and to track down a few customers. I even had a truly great discussion with James Bottomley, CTO of server virtualization at Odin, (and his exceptionally excellent bow tie) at the recent OpenStack conference in Vancouver.

Odin absolutely deserves to be on a Gartner Magic Quadrant. But it shouldn't be listed as yet another niche player in "virtualisation". It should be up in the upper right as one of the top providers of containerisation software available.

Virtuozzo is one of the best container systems available. It's perhaps worth a debate about Solaris's containerisation, but they are otherwise the best of what's out there.

Odin has the "visionary" part down pat and has been the software that underpins a huge chunk of the internet's "virtual private servers" available from various hosting companies for years.

If there was any company whose technology could be used to attempt to invalidate my distinction between containers and hypervisors it would be Odin's. They offer the ability to run more legacy style applications in containers – and with hypervisor-like features – than anyone else.

Despite this, Odin languishes in "niche players". An undeserved position for a company that is a leader in its market.

Virtualisation and containers serve different purposes. One day, maybe, they could be considered as one and the same. We've a lot of legacy software to deprecate – and a lot of features parity to achieve – before we reach that day. ®