"We build relationships, not just networks," - Sure, Convergence words to live by, but what kind of networks?
Network topology might seem like a dry subject for network architects and engineers to argue over into the late hours of the night whilst downing energy drinks and angrily brandishing routers in one another’s faces, but what does it mean to us?
Everything, that’s what.
Network topology defines many of the capabilities of a network and, critically for us at Convergence, has a huge influence on its resilience. Back in the bad-old-days, networks operated on a single line topology; that is, a point to point network in which each ‘node’ (the termination point of a single connection) was linked linearly to the next. This created a whole host of problems for the network: not only was it ridiculously vulnerable to damage (if one node or connection went down, so did the whole network) but it could be totally incapacitated by traffic. Needless to say, this type of network became rapidly outdated as information technology advanced. In order to improve reliability, protect the uptime of services and generally improve people’s lives, the Star network was created.
A star is born.
In a star network, peripheral nodes connect to a central node which controls communication between them. This allows the network to direct traffic, mitigating for the disastrous effects of high-usage common to an old-style connection. These networks were also more resilient; destruction of a single node would not destroy the capability of the entire network which increased downtime for the user and the bill for the provider. However, this network is not fool-proof, have you spotted the glaring flaw in this design?
Not all nodes are created equal.
That’s right, although some of the nodes in a star network can experience damage/destruction without effecting the performance of the entire network, there is one node that absolutely cannot do so; the central or ‘controller’ node. In a star topology, if the central node is destroyed there is no way the network can function, and all capability will be lost. So, whilst arguably the star network is less vulnerable than the old point-to-point one, by creating a central controller you’ve slapped a great big target on the head of one of the nodes, putting a bounty on it too tempting to resist. Any parties wishing to eliminate the network need only concentrate their efforts on the central node, a massive flaw in network security. An obvious solution to this problem however easily presents itself: why not have all the nodes connect to each other, decentralising control and therefore vulnerability? Wouldn’t it be possible to have a network where if one node failed, it would have no larger scale impact on the network at large? Why yes it would.
Mesh networks: the future of connectivity
Mesh networking topologies have been around since the 1980s and do what they say on the tin. Each node connects to many others, and there is no centralised controller to fail. The decentralised structure radically reduces network vulnerability and dramatically increases network resilience, however, despite this they have been slow to implement. This is because meshes remove the ‘middle man’ controller node (usually an internet provider) and therefore there has been little financial incentive to build them. Mesh networking is touted by many, including Qualcomm's Rahul Patel as the future of connectivity across the world and the movement toward them is growing.
At Convergence Group, we’re always looking forward to the future of connectivity. Our vision is clear, and we plan on being part of the network of tomorrow. We are driven, agile and no-nonsense. In an industry dominated by incumbents, we aim to be a business that will grow with the times, not stand against them. Our network is the network of the future, and we just can’t wait for tomorrow.
5 reasons why SD-WAN hasn’t lived up to the hype
How to implement WFH network strategies for the long-term