STRATEGIC ACQUISITIONS: Technology leaders need to adopt a comprehensive systems approach where each component, and therefore the system at large, is highly tuned to the task and optimised to make the most of data.
THE world is entering a new era of computing where trillions of interconnected things will generate more data than ever before.
By 2015, all these interconnected things will produce seven zettabytes of data around the world. For those counting, a zettabyte is a one followed by 21 zeroes.
This big data creates an increasingly complex computing environment, but also opens the door for incredible innovation - if we can find ways to use it to our advantage.
The future success of organisations relies on their ability to unlock greater intelligence, insight and visibility by tackling big data and better utilising structured data as well as unstructured data, such as videos, blogs and tweets.
But organisations struggle to meet growing demands on their computing systems to make the information useful.
Trying to keep up, they add more and more hardware to their existing systems, resulting in sprawling, unmanageable IT infrastructures. Such patched-up networks increase costs exponentially over time and prevents a holistic understanding of operations, forcing a cycle of poor choices.
The most common issue is that these are commodity computing systems, meaning they are built from basic parts, made for basic work. They do not recognise that not all computing tasks are created equal, and the same goes for computing systems.
Technology leaders need to adopt a comprehensive systems approach where each component, and therefore the system at large, is highly tuned to the task and optimised to make the most of data.
Computing systems are not unlike vehicles. There are many different types - pick-up trucks, sports cars, mini-vans, etc - but each serves a different purpose. No one has tried to sell a single type of vehicle since Henry Ford began production of his Model T vehicles more than 100 years ago.
One size does not fit all. Generic parts will only make up a generic system that cannot deliver smart results. Consider the following points when weighing which type of system can best help enterprises reach their potential.
Only efficient parts make up an efficient system. While each computing system as a whole needs to achieve one specific goal, each part can only do a portion of the work.
The better each part performs a particular task, the better the system will run altogether. In a commodity system, each part can only reach a certain level of efficiency, so many more parts have to work to get the job done.
On the other hand, if each part is specialised, it can perform one task at ultimate efficiency. Instead of harnessing the power of three standard cars to tow a mobile home, use one powerful truck.
By allocating resources as needed, a smarter system performs tasks quicker, using less power, and ultimately costing less money.
Short-term investments can have long-term costs. Shortsighted IT investments based on immediate cost-savings results in systems that cannot adapt to meet changing needs over time.
By making the right kind of investments, computing systems can have the capacity to benefit from innovations that come up tomorrow and many years from now. Otherwise, the expense of additions and upgrades will cost more in the end.
For example, cloud computing virtualises data, making it easier to store, share, and analyse. Cloud computing makes technology services available anytime and anywhere, benefits that will soon make it a common practice.
It is unlikely that a commodity system can deliver on these benefits because they require a highly sophisticated, highly tuned system. Imagine a car that only plays cassette tapes while the rest of the world is listening to satellite radio. It is about anticipating the future when designing your systems because you do not know what is ahead.
Quality of service relies on specialisation. Commodity systems lack the attributes important to specialised tasks. Would you use a standard economy car to travel off-road for emergency rescue or transport large amounts of hard cash? You would have to discretely add features across the board, which opens the door to errors.
To address specific work, a part needs specific accommodations, otherwise it can do many things, but none of them very well. Putting the flatbed of a truck on a sports car would result in a vehicle that is neither fast nor strong. Commodity parts, no matter how many you have, lack expertise and cannot provide the quality of service enterprises need, arguably the paramount requirement of any system.
To take advantage of all the possibilities of big data and this new computing era, organisations must build systems where each piece works in concert, each playing a differentiated role that feeds the larger operation. In doing so, they will create smarter computing systems that do more with less, use data to drive innovation and deliver results faster than ever.
Hemanth Kumar Kalikiri is country executive of the systems & technology group at IBM Malaysia