Given the quick obsolescence and the fast development in this space, keeping users happy is becoming the major task of the CIOs. Everyone wants the latest desktop or portable, irrespective of the applications they use it for. Chances are that CIOs face hardware dilemmas on two occasions—when buying new machines, or upgrading/discarding older ones.

This is the face of IT hardware as we know it, and in a way the troublesome side of it. Desktops and portables are the most visible components of an enterprise’s IT infrastructure. Also, the one that causes many a headache for the CIO who has to keep them up and running. And so, when we say think desktops/portables, it means think upgrades, think obsolescence, think manageability...

New purchases
While for portables, the choice is limited to a few players, you can do the jig while out shopping for desktops. Depending on your budget, you can stock your shopping cart with desktops whose price starts from Rs 5,000 (second-hand, bare-bone thin client), to assembled (less than 20K), to branded ones (25K upward). However, one has to be careful about the software that is present (read comes with the package).

Assembled usually costs slightly higher because of the rising levels of piracy and any installation of proprietary applications will cost that much more. Alternatively, you can look at the open-source basket and re-think your software requirements (see box on Desktop Software). Of course, this may not be a way to go while you are out looking for portables. Typically, a branded pachine would be the best bet and since they come with lot of bundled applications, it’s imperative to check on the company’s standardized platform.

To upgrade or to chuck…
Before you exclaim that this has been the age-old hardware dilemma, it might be time to consider this might be a situation courtesy Intel(!) Every few months, Intel comes out with a new processor as the availability of old models’ starts becoming an issue. Also, vendors start offering the same for both the corporate and home segments, irrespective of the applications intended to be used. The first issue to be dealt with is to predefine the bare-minimum specifications (specs). If you have the same, it is easier to follow the upgrade path clearly. (See box The Way We Do It). The next step is to have a clear roadmap of the future IT course of the company. For instance, if the company is planning to implement some new resource-hungry application in the near future, it would be best to junk existing machines and start planning for new ones.

The way out
The first step in managing your IT front-end is planning. This is imperative, given the high rate of obsolescence of these products. Planning would imply looking at the total cost of ownership—not just the cost of the desktop but also issues like standardization, cost of junking existing systems which fall below the benchmark, cost of setting up the machines in the office, configuring and networking a machine as well as maintaining it, upgrading the hardware and software, and dealing with user queries. These additional costs add up to several times the price of the machine. And the planning has to be an ongoing process.

For example, last year in Cyber Media (the publishers of Dataquest), for the editorial and other category, our CIO froze on Celeron or better, with 8-10 GB HDD, 64MB RAM , 15" color monitor , 4MB VRAM, CD-ROM, Mouse, NIC as the standardized platform. Given the obsolescence in the segment, the specs have since been upgraded to PIII/P4, 800 MHz—1.6 GHz, 128 MB, 20 GB HDD, 15" Colour, 52X CD-ROM, and all systems falling below these benchmarks are either being upgraded or scrapped.

On the notebooks’ side too, there’s no other way than proper planning. While the life of the machines can be increased marginally by upgrading, if the organization is planning to upgrade its OS platform, then upgrading might not be the answer at all, and would end up costing more in the longer run. In such an event, planning is the best possible solution. It is best to plan user needs and requirements on an annual basis and budget accordingly. It is best to limit the number of vendors to one or two for discounts and price benefits. Moreover, it is a good idea to have one or two machines as standbys, especially if you don’t have a standby machine clause in your agreement with the vendor.

YOGRAJ VARMA