Cisco Logo


Data Center and Cloud

Domain 7 in our Cisco Domain TenSM framework for data center transformation is what we call “Platform”.  More specifically, this term refers to the “software platform” upon which your business applications will run.  In short, this area is where we examine operating systems, databases and other types of middleware and help you figure out your strategy, architectural decisions and implementation plans in these areas, to help you drive a more successful cloud or data center project.  Let’s discuss this area in more detail.

First, though, if you are new to the Cisco Domain Ten, please check out my “Cisco Domain Ten: The Story So Far” summary blog I published recently.  Additionally, earlier this week, we ran a public webinar, where some of my colleagues in the Cisco Data Center and Cloud Services team gave their perspectives on Cisco Domain Ten.  If you missed this and their very practical insights, please do catch up on the Cisco Domain Ten webinar recording.

 

Cisco Domain Ten: Domain 7: Platform

Cisco Domain Ten: Domain 7: Platform

Let’s now consider a few quick observations and implications of Domain 7: Platform -- and again I’ll focus on implications particularly for cloud computing deployments (although most of my observations are pertinent to all areas of data center transformation).

My colleagues in the Cisco Data Center Services team, as we talk to customers across the world on their applications and platform strategy, confirm what is widely known: many data centers run with literally dozens (and sometimes way more than this) combinations of operating system versions, patch combinations and database versions and associated patches.  Can you imagine what this means when you need to troubleshoot issues.  It’s like running on shifting sand: you never know quite where you are, and you can be surprised by unexpected behaviours -- as a result of the variants in your underlying software platform components.  This increases your time to repair, your OPEX and your risk.

Not only that, if you allow many combinations of operating systems and middleware, your service catalog becomes very complex, your end users confused, and your provisioning support costs will escalate.

As an analogy, consider one of the low cost airlines -- for example South West Airlines in the United States or EasyJet in the UK (and I’m sure whatever country you are in, you have at least one similar type of airline).  They operate one or two types of aircraft, each with the same cabin layout, and all (airside) gate layouts will be standardized. As a result of the standardized platform they have put in place, when the come to launch a new route service, they don’t need to worry about whether pilot X can fly a plane type, and they don’t need to worry whether a particular plane can fit in at one of their served airports. Think of the business flexibility and efficiency this generates.  This is what you need for your data center software platform and applications!

When our consultants help our customers in this area, our mantra is “standardize, simplify and automate”.  From my previous blogs, this should not be a surprise.  In Domain 1: Facilities and Infrastructure, I made the case for standardization of hardware components as being a necessary pre-requisite for the scaling on demand expectation of cloud.

The same principles apply in the platform arena.  Take a disciplined approach.  Set a small number of “standards” for the platform layers in your data center.  Aim for one -- however let’s be realistic, I expect you’d have a handful of operating system/patch/database/middleware combination in your data center -- 3 or 5 “standard configurations” is way better than 30 or 50!  Manage your application suppliers and internal development teams to push these standards back on them.  Reduce your own variance -- don’t upgrade one server with a given patch, have a plan to apply this to all relevant servers where possible.

And give your staff a runway, not shifting sands, so that they can perform their provisioning and troubleshooting tasks on a standardised environment.  This will let you reduce risk, reduce OPEX, reduce mean time to repair, be more responsive to your end users, and make your cloud more successful -- all key benefits of the standardized approach which the Cisco Domain Ten framework advocates.

 

 

 

Comments Are Closed

  1. Return to Countries/Regions
  2. Return to Home
  1. All Data Center and Cloud
  2. Return to Home