Future network trends: wireless Wan will regroup

Until recently, a prediction that WiMax would gain traction in 2008 would have seemed pretty likely. But then, Sprint Nextel pulled its deal with US-based company Clearwire, which would have seen a country-wide network of towers supporting the wireless metropolitan area network technology.

Until recently, a prediction that WiMax would gain traction in 2008 would have seemed pretty likely. But then, Sprint Nextel pulled its deal with US-based company Clearwire, which would have seen a country-wide network of towers supporting the wireless metropolitan area network technology.

WiMax struggles on, but the utopian dream of ubiquitous access supported by advocates such as Intel seems to be fading, says Carmi Levy, senior vice-president in strategic consulting at Canadian analyst firm AR Communication.

In the meantime, Long Term Evolution (LTE) is gaining support from the industry. The update to the GSM-based UMTS 3G standard is a 4G competitor to WiFi, and it has the official support of the GSM Association. Verizon in the US has also announced that it will pilot a network based on the standard in 2008.

LTE supports cell sizes of tens of kilometres, and can provide download speeds of up to 100 Mbits per second. "This time last year, I do not think LTE was positioned as a viable alternative. It was not on the radar of most enterprise users or carriers. But then, WiMax got bogged down in ratification and suppliers did not come on board very much," says Levy.

Campuses and business parks may find themselves attracted to a technology such as LTE or WiMax, especially those that have delayed wired deployments in their own sites, preferring instead to wait for a wireless standard that requires less investment in equipment than a meshed WiFi network. But the 4G metropolotan area network standards must battle it out first, and Levy does not see them shaking out in 2008. The pilot activity and political jockeying will be fun to watch, but we will have to wait longer than a year for 4G to really capture the market's attention.

Wan optimisation will increase

Wan optimisation will be a growth area in 2008, says Neil Rickard, a research vice-president at Gartner. This product category, currently enjoying 10% to 20% market penetration, covers a multitude of different technologies. It focuses on improving the performance of Wans through techniques including the modification of existing protocols such as the common internet file system (CIFS). These protocols, which can often be unnecessarily chatty, do not cause major problems on a local area network because of the high-bandwidth, low-latency environment. However, on bandwidth-constrained wide area links, they become a problem. "Cisco is getting into that space, and so the area is undergoing tremendous growth," Rickard says.

Application delivery controllers will continue to hit the mainstream

Companies are rapidly realising that applications and networks do not exist independently of each other. If the applications are to function effectively, then the network must handle its traffic in an appropriate fashion for the application. This requires more intelligence in the core of the network, which must understand the nature of the traffic that it is carrying.

"Suppliers have different straplines in this space, but the idea is that the network is no longer just about moving IP and Ethernet packets from A to B. It will play a more active role in accelerating the application," Rickard says. In the case of application delivery controllers, that means taking some of the work traditionally done by the computing server and handling it independently. "It is about pulling stuff out of the application servers that would previously have been performing many of these functions pulling the mundane run-of-the-mill processing activities into the network."

An example of such a device is the Big IP appliance from F5 networks. A modular system, it can compress web traffic, prevent browsers from making repeated and unnecessary requests to the server, and will also implement traffic shaping in shared environments. "A server should not be setting up thousands of SSL sessions," Rickard points out.

VoIP will reach its tipping point

Voice over IP (VoIP) has been available for years, but for many companies, other projects have taken precedence. Jason Bremner, research director for infrastructure hardware at IDC Canada, believes that 2008 is the year when the market will finally see aggressive adoption of the technology among business customers. According to IDC's Canadian figures, almost one respondent in three said that their company would have completely installed or will be in the process of rolling out a voice over IP system during 2008. 35% could be either acquiring a system, piloting one, or at least considering the technology, leaving just a third of companies with no plans in that area.

Bremner hopes that this will be a stepping stone for more advanced services. "Realistically, you build the infrastructure first, and then you supply extra functionality on top," he says. "Cost is the primary reason that organisations jump in to IP telephony up front. They buy the IP telephony platform, they save some money, and then they take that and turn it into financing for some of the more sophisticated unified communications systems."

However, the numbers suggest that companies are much more cautious about unified communications. Just 15% will have installed or be rolling out a unified communications project next year. 25% would be acquiring or piloting the technology, leaving 60% with either no plans or no knowledge of their strategy in that area.

Green networking will become a central focus

If 2007 was the year of the green datacentre, then 2008 could be the year of the green network. Jon Collins, an analyst at Freeform Dynamics, says that as companies finally get to grips with introducing energy efficiency into datacentre environments, they will begin to look further afield for other potential areas of improvement. This will inevitably lead them out into the network infrastructure, creating opportunities for efficient equipment both in the core and at the end point. Energy demands on network equipment will increase as it is required to do more. As network speeds increase and devices ship with more intelligence, devices required to analyse packets at wire speed require more power.

Bremner's prediction of a tipping point for VoIP is particularly relevant here. Collins says that a rising demand for IP phones could create its own power-consumption issues. Many such devices, with their advanced processing capabilities, are essentially miniature web servers. They draw more power than a conventional telephone, which can be powered by the phone network alone. Driving energy efficiency into such devices will be a difficult challenge which companies may find themselves facing as they begin to embrace these technologies.

Application-level network services will increase in importance

Just as companies are realising the benefits of application-aware devices inside their core networks, so they are gradually embracing the idea of similar services in the cloud. Rickard says that application-level services on the internet will have a significant impact on the way that companies send traffic through public networks.

Akamai is one example of a company that is refining its services to appeal more to enterprise customers. The company, which caches content for faster delivery to consumers, has long supported the faster delivery of enterprise web applications. In October, it launched a service designed to accelerate the performance of non-web, IP-based applications behind the corporate firewall but over the public internet. Now, SSL sessions, Citrix-based application access, and VoIP can all be accelerated using the company's service.

This could herald a shift in network procurement policy from a capital expenditure to an operational expenditure model, as companies begin to realise the benefit of buying in services from others who really understand the subject area. "Technology offerings from service providers are maturing to the point where it is not second best to buying an appliance," Rickard says.

AR Partners' Levy says that such services will continue their extension into other areas such as network security. Google's acquisition of managed security firm Postini in 2007 indicates where the market value will lie. "There is less of a focus on buying a firewall at a set price," he says. "Instead, you work with a partner who has a core competency. Maybe they will give you the firewall as part of a broader service agreement. So in 2008 we'll continue a trend away from go-it-alone security to partnerships with third-party suppliers for delivery of subscription security services."

Gradual acceleration in the uptake of XML-aware appliances

XML routers have been around for years now, but according to Gartner's Rickard, they are still far from a mainstream product. He expects the gradual adoption of this technology to continue for the next two or three years before it becomes a mainstream category.

Companies that pursue XML processing in the network stand to reap some unique benefits, says Rickard. "If I had an application where I got a message every time someone sold something in a supermarket, I could aggregate all those messages. When the number of messages reached a certain threshold, I could go and tell a restocking application about it," he says. "And whenever another threshold had been reached, it could tell the marketing department's application. And you could put the rules into the network to control that."

Because network equipment that can process XML relies heavily on applications that can produce and consume these schemas, the development of this product area depends on the development of the application portfolio. This in turn is closely linked to web services and service oriented architecture concept, both of which have taken time to get off the ground because of the complexity of redevelopment. But 2008 will continue to see companies embrace the technology

2008 may not be a watershed year for any particular networking technology, but it will see the continuation of several exciting trends, along with a flurry of activity among some relatively new ones. Look to the coming year for an indication of larger developments, such as 4G, which will gain traction in the mid-term.

Read more on Wireless networking