As IT and network managers would know very well, improving network performance is simply a matter of moving the bottlenecks around. Not always, a network's bottleneck will be inadequate bandwidth. It may be underpowered servers, or out of date workstations.

It's no secret that network managers are under constant pressure to offer more services, to more users, while coping with new applications and technologies. In addition, they are asked to minimize the risk of downtime and failure, or rather maximize uptime, while lowering costs.

It's not uncommon that a cabling solution deployed today will be running out of capacity in the next five years.

Why? Because:

• Today's server will be tomorrow's desktop

• Every five years you can suffix a zero to your current Ethernet data rate

• Cost of bandwidth is dropping faster than packets on a poorly designed network

• Processing power will continue to double every 18–24 months while cost declines

• Terabytes of data moving and stored on today's networks will triple over the next five years

• High-performance computing, multi-site collaboration, and real-time streaming media will choke today's network infrastructure over the next five years

Hence the need to build in tolerance, by installing high-performance, high-quality systems while protecting against uncertain future developments in technologies, products, and applications that may require major (and costly) upgrades.

Such risks are managed only through a careful selection of systems with reliability, scalability and the future in mind. This is particularly true for the cabling infrastructure. Keeping the above in mind will help IT and network manager to have an infrastructure geared for the future.