Why Cisco's OTV Matters For The Private Cloud

We've all been hearing about it for a while, but I've just noticed that Cisco is getting more public in talking about their OTV (overlay transport virtualization) technology.

I consider this one of the key "enabling technologies" for many cloud models, and private clouds in particular.

And here's why ...

Disclaimer

I am not a network guy.  I don't even pretend to be one.

Also, I don't work for Cisco, although they're one of EMC's key partners.  So if I screw up this description, please use your greater expertise to help me out!

The Problem

If you disassemble the term "private cloud" into its constituent parts, the "private" part essentially means "under IT's control".  And the "cloud" part refers to pools of resources, dynamically consumed.

If we further take "IT resources" apart, we have mostly server, storage ... and network.

For each technology discipline, there's a clear technology agenda around what needs to be done to make private clouds an easier and more compelling proposition for enterprises and compatible service providers alike.

VMware recently added a bit more clarity to their contributions here.  EMC will be talking a great deal around what we're doing in our storage core competency, in addition to offerings in information protection, IT orchestration and information security.

Against that backdrop, Cisco's OTV is a big deal -- not only as a standalone offering from Cisco, but in the broader context of enterprise cloud-building.

It's My Network -- Or It Should Look Like It!

Consider the problem of multiple data centers where you'd like a management and operational model where it looks like a local LAN.  You'd like a uniform way of doing things, irrespective of whether the resources were in data center A, data center B or data center C.

IP addressing works the same.  Traffic management policies work the same.  Security works the same -- and so on.  What you'd like to do is to take disparate network domains, and have them look like one happy, integrated and consistent domain to the greatest degree possible.

Let's go a bit further.

Let's say that you want to do business with a compatible service provider, and federate some of their resources (infrastructure, platform or software) into your enterprise environment.  Again, you'd like the service provider network resources to appear as a logical extension of your own enterprise network to the greatest degree possible.

You can start to see the appeal of OTV, but there's more ...

Now, Let's Start Moving Stuff Around -- Dynamically

Those of us who work with this technology know that it won't be long before we can seriously consider creating pools of IT resources that span data centers -- irregardless of whether those data center resources are owned by IT, rented from a compatible service provider, or any combination.

Before long, you're going to want to dynamically pool over those resources -- whether it's to provide better protection and recovery for applications, or optimize costs, or provide a better user experience.

And you don't want your network topology getting in the way of moving a set of virtual machines -- and all of its relevant information -- from point A to point B.

What started as a "nice to have" becomes a "must have" -- there's no way you're going to fully exploit these resource pools unless you're presented with what looks like essentially a "flat" network.

There's More, Of Course

I'm sure there are lots of benefits associated with OTV.  And I'm sure Cisco is going to do a great job of explaining them to customers and the industry at large.

But, for those of us who are seriously engaged with the whole private cloud thing, this is a really big deal indeed ... :-)

Comments

[Charlie Gautreaux]

Chuck -

Nice summary. I believe it will take something like OTV to bridge the public/private cloud as well. Nice to see the wheels turning.

Charlie

[Ron Howell]

Chuck,

I believe your right about OTV and it's value. One thought, if OTV is to be implemented - it should be done within the Data Center or VBLOCK environment to be effectively quickly.

Why ? Because it can take carriers years to test and adopt a new technology such as OTV into their managed services.

If OTV just appears as a layer 2 feed to the Carrier Network, it can be implemented immediately and the benefits can be seen quickly.

food for thought,

Ron Howell
large carrier engineer
919-949-3513