22/09/11

By Ramin Attari, Managing Director Avaya Data Solutions EMEA

Unified Communications (UC) is all about integrating communication technologies like voice, video, email and text, so they work together to optimise business processes, access to information and speed up decision-making. When done properly, UC can help fuel business growth by making employees more productive and elevating customer service above competitors.

Today’s climate mandates fiscal prudence, so we need to be mindful that times are hard and any incremental spend in IT infrastructure must be justified. A UC framework based on SIP (Session Initiation Protocol) can easily resolve this dilemma as it allows organisations to extend the life of most elements of their existing infrastructure (through SIP enablement) without a significant additional investment. In addition, SIP can provide new functionality in the form of improved presence and other beneficial features.

However, a UC framework is only as strong as the underlying infrastructure upon which it sits. Voice, data and video services must also rest on a common infrastructure to deliver the maximum benefit to the business in terms of cost, management and simplicity. It is possible to host and manage separate networks for each of these elements, but it is expensive and can be difficult to scale, so one network is almost always better than three. It is though absolutely essential that that the supporting data network can cope with the increased bandwidth demands placed upon it by the evolving requirements of the business.

In times gone by, the demand for more bandwidth from new applications resulted in (almost automatic) network upgrades, but this approach is much more difficult to justify today. We need an alternative strategy, one that combines the ability to drive the business forward with the need to maximise the value of network assets and maintain a balanced budget.

One way to reconcile the need for evolution with cost is to take a fresh look at the architecture of the underlying network to see if a different model will not only deliver the performance, resiliency and security needed, but deliver it at a lower cost. Absolutely, the data network must be rock solid and able to detect, react and recover from failures before users notice, but it doesn’t have to cost the earth.

The traditional way network resiliency was achieved was by having two of everything - one device or link to deliver the service and another as a backup. To put this into perspective, this is like having one highway to carry all the traffic, and a back-up highway that is only used if the first one cannot be used for some reason. In the event of failure, the backup takes over and 100% service levels are maintained. The problem with this approach is that having two of everything is again very expensive. It is also wasteful on resources such as electricity and is arguably unnecessary since failures are very infrequent. It is much more cost effective to derive resiliency through an active-active model that balances the total load across two or more smaller devices concurrently. Under normal conditions everything is active and capacity is the same as the traditional model. In the unlikely event of a failure, service is maintained, just at a reduced level. By running the applications over both “roads”, businesses can either double their capacity or halve the cost with very little effort.

UC and other converged services help communications technologies work together, but it also increases the amount of traffic running across the network. In extreme cases this can cause congestion which results in decreased voice and video quality and slower application response times. To prevent this, the network must have the ability to prioritise certain types of data during their journey from sender to receiver and ensure a Quality of Service (QoS) for the application. For example, streaming media applications, such as voice and video, need a high level of priority because they are delay sensitive (we hear and see the impact of a delay immediately). Applications such as email, IM, file transfer and web browsing can have a lower level of priority, not because they are less important to the business, but because they are delay tolerant (their passage through the network could be delayed without causing much impact to the business).

In an ideal scenario there is always enough bandwidth, so congestion never occurs, and the prioritisation of traffic never needs to be enforced. Even in these scenarios, QoS can be a benefit, since it can also be used to ensure the network is used to capacity, reducing the need for incremental spend on upgrades. A well implemented QoS policy could delay the need for a network upgrade by many months. It is also a good solution for dealing with temporary increases in demand such as during a busy season or time of day. QoS can be an extremely useful tool to allow the deployment of new applications or reduce bandwidth costs when budgets are stretched.

UC stands on a foundation of data, which has the potential to deliver many benefits to business in productivity and communications. The challenge is managing it properly, so it can support businesses as they evolve and change.