Tiered Infrastructure Architecture

Thanks to the recession, CIOs are busy but not to support the growing demands of business, but instead identifying how and where to save costs and bring sustained value. When the going is good, it seems to pay better to throw more hardware, software and labor at any business requirement because business is growing and someone will pay for it. Studying delta savings and incremental benefits are more of overheads, where the cost of managing them seems to outweigh the benefits accrued. Not any more in these times!

One of the things I have consistently observed is the lack of a consistent approach to a tiered infrastructure architecture. IT infrastructure is mostly set up as chunks of monolithic or patchy systems with no stratification or gradation of the infrastructure or the service quality provided on top of it. While many banking and other enterprises do have "A Class" infrastructure or service for "A Class" business applications, the rest of the applications and their underlying infrastructure is most often a mix of systems acquired over time. It is often not possible to say that we have three kinds of storage infrastructure --- A, B and C which caters to different kinds of storage requirements in terms of these metrics 1. ..... 2. .... and .. 10. ...

An efficient environment should have an enterprise divide its IT infrastructure into two or three grades. Each component (hardware, software, labor) should fall into one of these buckets. It may not be possible to have an air tight environment with no overlap but having at least 80% discipline from the current 10% (or none) will result in cost efficiencies and better manageability. It will also ensure IT is more responsive and participative in growth --- next time a project estimate is made for (say..) setting up a new back office, the project cost would be more reflective of how critical that operation is to business. IT team would first ask how critical that environment is, operational requirements, redundancy requirements etc. for all aspects of IT environment to plan for an optimal set-up instead of an one-quality-fits-all approach where that "one quality" is often an expensive service requirement.

Where to Find Dollars in an IT environment

Where are those hidden dollars, those waiting-to-be-unchained values in an IT department? There are numerous nooks and corners where there are hidden opportunities to save. Bad times like these throw lot of opportunities to bring cost efficiency. The role of an outsourced provider is so much more critical in such situations. Of course there has to be the right level of motivation as part of the contract to motivate the provider to help the customer find oppportunities to save.

Outsourced providers present themselves with this opportunity by having a view across their engagements of common such opportunities, plus themselves running the operations of a particular customer, having the requisite information to enable such cost saving.

There can be numerous such opportunities but here is a start to list some of them. Can't cover all in this post but will take a beginning:

• Unused licenses lying in the enterprise (and many find themselves getting budgets to buy more of the same for new requirements). These are more for end user software like MS Project, Visio etc.
• Underutilized hardware in parts or whole
• Spare network bandwidth in tier-2 or tier-3 of the networks off the central network topology which is more closely monitored
• Groups of support teams (in-house and or outsourced teams) providing services very similar in nature as part of different businesses
• Multiple data centers or server rooms
• De-centralized environments making it necessary to have lumpy teams across the country or globe doing pretty much the same stuff.
• De-centralized purchasing of IT assets

This is just the beginning and more are sure to follow.