I’ve recently been given the opportunity to write a blog for Global Crossing – a much larger audience in which to stand up on my soapbox and say all the things that have been bugging me about the Ethernet industry for years. But, I'm not going to do that.  Well, not today.

If you have read my blogs on http://www.carrierethernet.us.com or http://vpls.us you will recognize my posts – But for those of you who haven't read my posts, I am looking forward to your comments.

Anyway, that's not the reason that I came to talk to you today

The last conversation I had on my blog surrounded the newest carrier hotel / Ethernet connection point that CENX and Equinix have been presenting. They are providing the ability to connect with anybody that they can interconnect with to themselves, and in return they will provide the ability to connect with everybody else at that interconnection point. Driving costs and service provider profits down by creating a bottom feeding situation. Have you ever heard of the “Name your own price strategy” a couple of the mortgage vendors and a couple of the airline cheap ticket web sites were providing this service. What I think they envision is where a service provider has to bid on a customers connection and compete against all of our competition for price. The problem with this situation is until it becomes a commodity, or an even playing field, it’s not an applicable approach. 

When the offer we have is vastly superior to everybody else, say bursting, or quality of service or service multiplexing or a whole host of other things, it hurts the value added proposition that Global Crossing and some of the other Service Providers offer.

But that's beyond the scope of this blog

This blog post was to just mention that what Cenx and Equinix don't realize is that we have been stealthily working towards a similar model for the last 2 years. Hooking up carriers, getting interconnection strategies in place, figuring out the offer of almost EVERY Ethernet carrier in the world and preparing ourselves for WORLD DOMINATION. AHAHAHAHAHA (Best evil genius voice) Of course I'm sort of teasing. We weren't really trying for world domination, nor are we trying to compete against Cenx or Equinix, we simply aren't in that market.

Our intent is/was to:

• Increase our limited last mile access. We have some access but in lots of cases we don't have the last mile. We have to rent it from somebody or pull fiber at a huge expense.
• We wanted to get into markets that we don't have either a network or a partner
• Reduce or own internal cost of access. Having multiple providers inside of a region allows us to price multiple options versus quality and get a better rate.
• Offer the service. It's not really a service if our customers cannot get to us.

Since both models are accomplishing the same feat with drastically different approaches to the problem, I’m proposing two additional terms to differentiate our offering from theirs. ENNI-SM, ENNI-DM. Or in other words, Ethernet Network to Network Interconnections in two flavors, Sparse mode and Dense Mode.

• ENNI-DM: this is an interconnection model that provides lots of interconnections at a single location. 3+ connections that allow lots of local loops to be terminated in a single location. This is typical of a local meet me room or of a interconnection point where we can either be physically or logically switched to another service provider through a third party.

The diagram above basically show how Cenx is offering a location (3) in which they are providing GE interfaces to their gear and then offering interconnections between you and everybody else who buys into this idea.  Point A and Point B could be in multiple locations which may or may not may make this the most applicable design.

• ENNI-SM: This interconnection model is a distributed model of interconnections. The Third party becomes the transport portion of the equation. So instead of dense amounts of connections into a single location the service provider provides the interconnection in a geographical dispersed method.

Why would a customer want a different model?  It really becomes a set of several questions:

• How does the customer wants to have the traffic carried. Is it cheaper to carry it across their own backbone or put it out to a third party?
• Does the customer have footprint in the remote location? Or does it have to traverse yet another Service Provider?

So what did Global Crossing do or how did this come about? Well it was really a side effect of our direction.  What we did was in lack of an established way in which to build out our access network, we did what service providers have done for years. We started hooking up other networks to our network. We've done that through class IV switches, we've done that with IP networks, we've done it with Frame Relay and with ATM. So it wasn't a stretch, and basically fit into our standard model. Basically the ENNI-SM method described above.

So, you see where Cenx is providing a box with a single or 3 interconnection points? Well, Global Crossing is offering 50+ at the time of this writing. 50 locations and 50 plus providers. So what we've done is basically turned their box into a cloud. Is this "Cloud Computing" at it's finest?  Turning a resource into a cloud? Isn't that what cloud computing is all about? Perhaps not, I'm the Ethernet transport guy not the cloud computing guy.

Questions on this or other Ethernet Topics? 

I can get as detailed as you need in the discussion of our product or in the theory of VPLS / Ethernet Transport and basically anything related to Ethernet.  Your different topics will start different posts.  Is there anything you want me to discuss?