Translate this website:
Search this website:

CitrixDesktop VirtualizationHyper-VI/O VirtualizationNetwork VirtualizationRedHatServer VirtualizationStorage VirtualizationVMware

One Size Does Not Fit All

By Mike Williams, Distinguished Engineer of Systems & Technology Group at IBM.


Date: 23 May 2011

Despite best intentions and even with the most informed experts by our side, data centre and respective business managers face a vicious cycle of numerous infrastructure challenges. Some have un-trusted data. This means we’re always guessing, making decisions based on incomplete data. Big ideas are seen as risky, and small decisions aren’t optimized. Some have sprawling IT. This means everything costs more. Every IT investment leads to more sprawl, driving up infrastructure and management costs. Some have inflexible IT. This means systems and business operations are often reactive. Inflexible infrastructure limits system integration across silos and responsiveness to customer demands. Many face most of these challenges and some face all of the above.

There is no secret recipe to fixing these problems. The simple solution is to make the most informed decisions and plan for an IT roadmap that extracts long-term economics and business value.

Some advocates profess that the only way to reduce the cost and complexity of computing is to adopt one kind of computer for the data centre. Organisations that rely on a single computing platform or a single delivery model do so at their own peril, because not all computing tasks are created equal, and the same goes for computing systems.

Consider computing systems as you do a vehicle. There are many different types – pick-up trucks, sports cars, mini-vans, etc. Each serves a different purpose. We can’t choose one car, get rid of the rest, and think it will fit everyone’s needs. Diversity allows us to choose the most efficient and effective car to complete the desired task. One type of car for everything would also sacrifice performance. Creating a sports car with the flatbed of a truck would result in a car that poorly performs either task.

Here are five points to consider why “one size does not fit all” in efforts to drive a more efficient, smarter computing environment.

1. Look under the hood and all the piece parts.
At its foundation, a commodity system essentially means that the supplier has bundled various components, melded together with basic piece parts. This system is inexpensive yet shortsighted. Innovation and adaptability are non-existent because production cost-savings take precedence. For example, consider trying to create a world-class car with bits and pieces from cheap, used models. Also, this commodity-based system is rigid, capping utilization rates. As a result, efficiencies reach a threshold and critical operations are stalled because of poor availability features. Imagine having invested in a standard family sedan and then learn that your family has grown. You now need to transport your children’s friends as well. Or, your needs continue to vary like transporting large furniture as the kids move to attend a university.

Before long, IT is entrapped in endless cycles of rigid, commodity systems that were never designed for ultimate flexibility. Commodity systems can become dangerous traps for server sprawl and inflexibility. In simpler terms, it’s about “future-proofing your systems because you don’t know what’s ahead.”

2. Keep an eagle’s eye on security.
Another area of neglect common among commodity systems is security. Commodity options often lack proper business-specific security mechanisms, forcing IT to slow down the business cycle to allow time for thorough security integration. There’s also the need to find and hire security experts willing to work on an ad hoc basis.

3. Pay attention to quality of service
Built with basic functionalities and from common components, commodity systems lack the resilience and serviceability of a specialized system. Critical attributes like redundancy, single points of failure, zero downtime and performance are non-existent in commodity systems. If your shipping business depends on on-time delivery of goods, no basic economical truck will do. You’ll need one that delivers traffic alerts or weather conditions because your customers simply demand their shipments as promised - no excuses. Some may even argue that qualities of service requirements are most paramount.

4. Use systems innovations to drive business value.
As IT budgets continue to remain flat while business demands skyrocket, one area of technology that has become extremely popular is cloud computing. Anywhere, any time access to key technology services has only increased the need to run an innovative system. It’s important to ask how and if a particular system can deliver on the bankable attributes associated with clouds, like consumable Web-delivered services, flexible pricing, rapid provisioning, elastic scaling, on demand features, account metering and billing, and advanced virtualization. These are all elements that require a highly sophisticated, highly tuned system.

5. Ask how the system meets different business scenarios and requirements.
Many computing tasks are variable and many computing systems need dynamic capabilities to adapt to meet new demands. For example, a transactional system that serves thousands of online users, 24/7, has different needs than an analytics system that has a very small number of users but they require complex queries. Systems must be tuned to the specific task or workload, a feat that few commodity systems can achieve. An average family four-door car does not have the finely tuned engine to compete in a race nor the pulling power built into its bumper to pull a large boat.

Whether you’re in the business of saving lives through powerful computing or predicting fraud to safeguard your organization, taking a stringent review of every system option is critical. You would not blindly buy the most economical car when the future needs of your family are unpredictable. And whether you decide on a commodity or specialized system, we all can agree that smarter computing is about doing more with less, unlocking insights from data to drive profits, and delivering it in a way that services can be turned on at a moment’s notice.


« Previous article

Next article »

Tags: Server Virtualization

Related News

28 Sep 2015 | Server Virtualization

22 Sep 2015 | Server Virtualization

22 Sep 2015 | Server Virtualization

17 Sep 2015 | Server Virtualization

Read more News »

Latest IT jobs from leading companies.


Click here for full listings»