As the state of the UK economy improves, many organisations will begin to look to their IT in the quest for a competitive edge. Over the past decade, and prior to the global recession, the investment banking sector looked to “more, faster, better” IT as a way to gain an advantage over the competition. Not any more. We look at how the playing field has changed for large organisations and some of the technologies being used to make the most of what you’ve got.

If you’ve got it, don’t flaunt it

During the height of the economic boom and after the dot com bubble had burst, the attitude in the investment banking sector was very much that, when the skills of the software developers, systems engineers and enterprise architects had been exhausted, the only way to improve performance was to buy more powerful hardware. Rumour has it that, as their nearest rival had started to process complex mathematical calculations, predicting trends in the price of credit derivatives quicker than them, a couple of traders in one firm simply wrote personal cheques for close to half a million pounds each, to buy the latest and greatest enterprise server, thus stealing a march on their adversaries. It worked for a while, until a new, more powerful server came out and the whole cycle started over. Those days are all but gone. Money spent on IT has been curtailed because there isn’t any, and the focus has shifted to more intelligent use of the hardware in place and to the careful sharing of resources through virtualisation, cloud* based and high powered, distributed computing power (also known as ‘grid’ compute farms).

Back to basics

Global organisations are making use of all that ‘tin’ they spent so much on in the past, and improving the way they provide access to computing resources, not only on the servers in data centres doing all the most important information processing, but also to end users at the desktop.  Servers, once used by distinct business groups are now being shared with others, so as to ensure 100% of the available CPU power being paid for is being used. The same can be said for disk space, and networking capacity. Using virtualisation technologies, such as Microsoft’s HyperV, or VMware, and the right tools for physical to virtual migration (P2V) new generations of virtual servers are being commissioned. Through companies such as VKernel, who provide tools for finding underutilised resources in the virtual estate,  hardware can be utilised more effectively.

Large firms are even looking to the lowly desktop PC to harness all the spare CPU cycles available to them. Open source projects, such as Folding@Home and SETI@home have produced distributed computing applications, started alongside the computer’s screensave. They are designed to run jobs on desktop PC’s at night with the resulting data being aggregated centrally, similar to grid compute farms. Large organisations are looking at this technology as a serious option to get the most from the hardware they have. But, in order to fully harness this spare compute power, they need effective configuration management in place; imagine the problems with two different versions of the software installed on the PC.

Of course, in many cases, the humble PC is now being replaced by the thin client and the web browser, email, word processing and spreadsheet software that once needed to be installed and maintained locally on each user’s desktop computer (no mean feat, if your firm has 300,000 users to look after) can now run ’somewhere in the cloud’. Advances in the way graphics can be transmitted over a computer network mean that, even though a Windows (or Linux) desktop is being run on a server somewhere in a far off datacentre, the user experience can be as though it were running on the PC in front of them. This technology is advancing quickly, and large firms need to ensure a strong awareness of what is available and how it would fit with their organisation.

Where have I seen that before?

Remote desktop technology is not new: Sun Secure Global Desktop from Sun Microsystems has been available for several years, providing very lightweight communications mechanisms, carrying bandwidth-intensive graphics traffic across wide area networks; it was used on offshore oil rigs to present data to the rig operators, without the need for local computing power to carry out the complex calculations used by seismologists to prospect for new oil. More recent advances, such as HDX from Citrix (who have also been providing remote desktop software for years) or PCOIP from Teradici now mean that highly graphically intensive applications, to the extent required by CAD and multimedia software, can be run from servers in data centres around the globe, with the end users in another location altogether.†

It should be noted that, contrary to the perceived benefit of moving to thin client technologies, the cost can actually increase significantly, with the implementation of products such as PCOIP. It works by effectively moving the body of the PC into the datacenter where specialised data compression is carried out before being transmitted across the network, and the data decompressed at the desktop, via proprietory hardware in the thin client device. So, although there are effective savings to be had in power consumption on the desktop, they use more energy in the datacentre.

The proof of the pudding..

Only time will tell who proves to be the earliest adopter, or the fastest trailblazer, with all this new technology floating around in the cloud. Will the banks, hedge funds and insurance firms be the first to adopt these new technologies? That is difficult to predict; many are still struggling to squeeze any more out of their CFOs for new IT infrastructure and projects. In organisations with 100,000+ employees, even a move to a new, cheaper method of providing desktop IT through the private cloud has to be controlled through a large scale project, involving programme and project managers, architects, engineers and systems administrators. Done properly, cost savings can be had from increased utilisation through virtualisation, as well as improved time-to-market of desktop services.

* There are many definitions of cloud computing circulating at present. For the purposes of this article, we’ll define the cloud as pretty much anywhere in the world, other than your desk, or a little room in the back of your office.
† Sadly, there is one thing we cannot overcome at present and that’s the speed of light in a vacuum. It is what causes latency (delays in the round trip time for two computers to communicate to each other over very long distances). Bangkok to London, for example, takes 0.06 seconds to make each 20,000km round trip journey. And that’s assuming high speed, fibre optic cable connects both ends all the way, and the networking hardware is up to the job. So, there is a limit on how far apart the end user can practicably be from the computers their application is running on, if a real-time application is what is required.