The ongoing pressure for businesses to reduce costs and increase output is often felt most by the IT department, with technological implementations expected to improve efficiency and productivity. More and more, businesses are also looking for greater yield from their existing technologies, so many IT departments are redesigning their infrastructures to increase efficiency and improve effectiveness.

 

Developments in virtualization, storage and networking technologies have resulted in a high proportion of organizations adopting a centralized, consolidated IT infrastructure. Rather than the more traditional model of IT resources being spread out around the organization in silos, many businesses are moving IT functions to just a few (often one) main data centers, where the technology is more easily managed and maintained.

 

This consolidated infrastructure model is only possible as a result of the strides made in business-quality networks over the last ten years or so. Even in organizations with geographically dispersed offices or branches, applications can be held centrally and accessed by all areas of the business using a wide-area network (WAN). The hardware deployed by the business can often be consolidated to reduce the operational cost burden on the business, and controlling and supporting the technology becomes more straightforward. In addition, uniting the IT workforce means staff can be better employed.

 

The benefits of transitioning to this kind of structure seem apparent to many businesses; however, there are generally a great number of complications associated with any restructuring project. Many of the fears that arise around centralized IT systems relate to reliability issues. Organizations are concerned about problems with hardware or the physical space it is housed in, resulting in a loss of access to all systems, negatively impacting productivity across the entire business. Likewise, the network is another source of concern as the effectiveness of a centralized infrastructure relies entirely on application traffic traveling over the network quickly and smoothly.

 

Business networks are coming under greater pressure as the reliance on them underpinning operations grows. The take-up of Internet-based communications increases this strain as very sensitive real-time voice and video applications are added to existing email and messaging tools. This traffic has to compete with a range of other types of applications being accessed via the network, from transactional systems such as ERP or CRM to email, file transfers and backups. Each traffic type has its own distinct way of behaving and reacts to congestion on the network differently.

 

Consolidation of servers and applications also put extra strain on the network, and as the amount of traffic on the WAN increases, applications become more likely to experience performance problems such as delay, jitter or loss of data packets. According to research commissioned by Ipanema (News - Alert) in July 2007, application performance issues are costing UK enterprises more than £500,000 per year due mainly to reduced productivity, lost sales and diminished customer confidence. It doesn’t reflect well on the business if the CEO’s important video conference with Japan suffers due to congestion on the network, and the blame for this kind of incident is generally laid squarely at the IT department’s door.

 

The research revealed that businesses most commonly respond by increasing spending on bandwidth; however, while 57 percent admit resorting to this tactic, 75 percent of all businesses polled complained that bandwidth upgrades rarely solve the application performance problems they were experiencing. As a result, a considerable 93 percent of respondents said they are looking for greater visibility and control of traffic over the WAN. Evidently, continually increasing the network pipe at great cost is not the solution; rather, IT departments need to be able to manage the traffic dynamically, prioritizing applications based on their sensitivities and business-criticality.

 

To achieve this, businesses need to be able to examine what’s actually happening on the network, and then decide on the most appropriate response to ensure that business needs are being met. In a large number of cases, performance problems will be caused by the congestion of network traffic peaking at certain times, such as during the lunch hour when many staff may surf the net or shop online. Most of the time, the network performs without a problem, and there is no need to increase bandwidth. Without network monitoring tools in place, IT staff will not be able to see this is the case, making it much harder for them to find a solution that improves end-user experience and limits expenditure.

 

Without being able to see exactly what’s happening on the network, finding a solution that answers both requirements is impossible. As such, IT directors will benefit most from being able to monitor network performance in a way that not only highlights what’s going on, but also allows for adjustment and improvement to be made so that costs are kept down and performance of critical applications can be guaranteed in times of congestion.

 

A system that monitors the network, recognizing different types of traffic and examining the impact they are having on performance provides the network manager with information on areas where there may be unnecessary duplication, or traffic could be staggered (for example, anti-virus update downloads). This is a great starting point for improving the performance of the critical applications over the network.

 

The Holy Grail, though, must surely be a system that not only monitors how the network is performing, but also recognizes applications deemed most critical to the business. By dynamically prioritizing this traffic during periods of high demand, problems in performance that may otherwise have occurred can be avoided.

 

With such a system in place, network managers would be able to understand how applications are performing over the network, and build on this by providing performance guarantees to the end-users of business-critical applications without paying for masses of bandwidth that go largely unused.  In this way, networks in a centralized infrastructure can be relied upon to serve the needs of the entire business.