I previously wrote about next-generation networks, which have been identified as possibly being the start of the hybrid phase of a converged communications infrastructure.
I will now expand on an equally important, yet lesser-known subject, which is referred to as network-aware applications or next-generation applications.
Network-aware applications refer to applications that can, under certain conditions, detect if the network is congested or if the delays on the network are too long or whether there is a problem in the overall network.
The ability to detect these shortcomings is critical to the network and consequently there have been numerous attempts at developing hardware and software to address these limitations. Packeteer, Peribit and Expand are all examples of such solutions that have a bearing on user experience and contribute to quality of service (QoS).
What are network-aware applications?
Let`s assume we want to send a large volume of data from one point of the network to another and we only have a 33.6Kb line. The application should be able to detect the slow speed network connection and compress the data so as to reduce load and improve performance.
Similarly, if one were in the middle of a voice conversation and the delay over the network became unacceptable, the mechanism would adjust the priority of the voice packets. It would also have the intelligence to seek out alternate paths that would bring the voice conversation back within acceptable limits.
TCP/IP runs each individual application as a service on a service port. This helps to distinguish between applications such as http for Web browsing or RTP for voice. The TCP/IP protocol connects via sockets and it is at this level that the network-aware application intelligence is being explored.
Once applications are network-aware, the next step is to get the applications and operating systems to become security-, processor- and memory-aware.
The combination of these is then referred to as a self-healing network - a concept that has been on the cards for years now. From a service perspective and operational expenditure point of view, there is no doubt that having network-aware applications will benefit any ICT infrastructure. It also means that companies can save on costs as fewer operators and technicians are needed to maintain the network.
These mechanisms bring reliability and sustained acceptable service to the organisation and, of course, the network. However, the reality is that all of this is not new and organisations that are first to move in this direction will be the companies that survive.
Taking the lead
Once applications are network-aware, the next step is to get the applications and operating systems to become security-, processor- and memory-aware.
Andy Brauer, chief technology executive, Business Connexion`s Networks Competency.
It is no coincidence then that the open source community is the one that seems to be taking the lead in this area. The reason for this can be attributed to the fact that the play station generation is growing up and they are driving the need for the self-healing network.
IBM claims to have built a machine with the capability of 25% of a human brain in 2004. Further research indicates that we will reach singularity by the year 2035. This means that man and machine will have equal intelligence by then. The bottom line here is to not stagnate, but keep on learning.
Another term that we need to bear in mind when speaking about network-aware applications is traffic engineering. Traffic engineering is achieved by moving large volumes of user traffic along predetermined paths that traverse specific nodes in the service provider`s network.
A worthy description for traffic engineering can be found in a white paper published by Juniper Networks in the US. The white paper defines traffic engineering as being concerned with the "task of mapping traffic flows to the physical network topology. Specifically, it provides the ability to move traffic flows away from the shortest path calculated by the Interior Gateway Protocol and onto a less congested path.
"The purpose of traffic engineering is to balance the traffic load on the various links, routers and switches in the network so that none of these components are over-utilised or under-utilised. Traffic engineering allows an Internet service provider to fully exploit its network infrastructure," states the white paper.
While traffic engineering refers to congested routes, load balancing and alternate paths, it does not address individual application issues and performance, which is the task of QoS.
Share