Thursday, May 15, 2014

Local Clouds and The Coming Death of Legacy Stacks in the Cloud


Let me tell you a little secret about the “cloud.” It’s that right now in the enterprise, it’s a local game.

A few options come to mind when we think of enterprise, like the predominate force of AWS in test and development environments. Then you have providers like, HP-ES, IBM/SL, RackSpace, Teremark, and Google that are all trying to play enterprise production load catch-up.

For the meantime, I think it’s safe to call the enterprise private cloud a local game.

Take KIO Networks in Mexico City, or LGCNS in South Korea, or T-Systems in Germany; the strength that these local service providers have is that they are entrenched into the local economy. They are tied in with the governmental entities either with contracts, investment tax incentives, or in some cases, board relationships.  Governments encourage these types of business, as they are clean, provide local jobs, and are a good high-tech face for the country.

The key characteristics of these service providers are that they have the local connections and the P&L margin expectations for the long-tail economics of a service provider business.

In the case of KIO Networks, their margins are so tight that they build their data centers in cooler zones in Mexico City or bury them into the side of a mountain; because not running their chillers for several months out of the year is a competitive differentiator. Coming from a world of plump enterprise software/systems deployments, margins like this seemed like a foreign concept to me.

Just like our friends at AWS, service providers don’t write the books, they just sell them and are perfectly happy living with the retail economics. 

But in this never ending quest for margin, service providers need to standardize on control layers to manage each plane, like, computing, networking, storing, provisioning, and managing chargeback across these heterogeneous, at times customer dictated, and other times, commodity resource pools is vital. 
This long-term platform migration, which is primarily driven by the evolution of the service provider, is what’s posing a sea changing threat to the legacy of major profit pools from the likes of: IBM, HP, EMC, NetApp, etc.

In the early stages of a customer’s journey into the private/public cloud, they generally dictate the same legacy platforms they have run on for decades with these environments lifted and shifted into the services provider data center. In other words, let’s move my expensive proprietary boxes off of my data center floor and onto yours. 

This is precisely where the market is now.  But this phenomenon is just a hosting/colo way station on the way to the true public cloud.

Going forward, the confluence of software driven reliability, fault tolerance, and compliance are being delivered on top of commodity infrastructure selected by the service providers will be a way of life.   

Service provider’s margins will never tolerate the proprietary stacks of today.  

These business models are too far out of synch. AWS’s cloud doesn’t run on proprietary gear, why should yours?

No comments:

Post a Comment