Dual Core Network Design – Dual core design, also known as dual plane or disjoint plane topologies, refers to a highly redundant network chosen by companies whose main objective is to improve the resiliency of their network. Created using different data planes, dual core design is implemented by companies that receive the service from the different service providers. Put simply, big companies use dual core design in order to improve their network. Institutions that generally use this design are found in Europe; they include banks, hospitals, and other financial institutions. What’s more, some companies outside Europe use this design.
The links – passing through same fiber conduit, building, town, or city – are identified as Shared Risk Link Group (SRLG) since they share the same fate if there are any technical glitches. It is pertinent to carefully identify SRLG links between the providers. And if there are shared links, diverge links should be demanded.
The figure above depicts the dual core design topology. The control plane (routing protocols, management traffic, and any inter-networking node traffic) and data plane (user and application traffic) are separated from each planes. Separate IGP and separate BGP are used to protect one plane from failure at the other plane. For example, OSPF is on one plane and IS-IS is on the other plane, both of which are used for routing protocol.
Control of the core networks with the layer-3 VPNs depends entirely on the service provider even though two layer-3 services from two different service providers are not a dual core design. To name a design a dual core design, the company itself should control the core. If an enterprise receives layer-2 service from one provider and uses internet from another provider, dual core design can be created. With layer-2 service, a company can control its core, create a VPN, and control its core network using the internet. For Dual core design, some companies use different vendor equipment at the two planes. Also, there is no routing redistribution between the cores.
Careful filtering along the edge is critical for dual core topologies, as you will see below.
In the picture shown above, routers A, B, C, D, and E are one core; routers P, Q, R, S, and T are another core. These two cores are unconnected in any way –there is no way to forward traffic from router D to Router S, for instance, without passing through either router Y or Z. Routers Y and Z, however, are configured so that routing information is not shared between the two cores. Another reason for the configuration is to prevent traffic from passing through the cores. More importantly, filtering at router Y and router Z is very critical in order to prevent routing between the planes.
Dual core design may require two different teams to operate a network. Ring or partially mesh topology can be used in each core. Thus, convergence of dual plane topologies depends on each underlying topology. On the one hand, one network core is ring; on the other hand, the core can be meshed partially. Given the features of the network cores, there is an opportunity for segmentation. While low latency application is sent through partially-mesh plane, high bandwidth application is sent through Ring topology plane. Undoubtedly, these processes guarantees better quality of service.
An important benefit of dual core network design is service deployment. For instance, if a new MPLS traffic engineering service is deployed at only one core and if this new service causes network failure or service degradation – critical network traffic can be quickly moved to an unmodified plane along the network edge. What’s more, service can be restored without the new service being out of production, giving network engineers more time to troubleshoot the problem.
As long as fast service adoption, high availability (5x9s and more), and business continuity are very critical, dual plane is an attractive design. Deploying different vendor equipment to dissimilar cores prevents vendor lock-in; it also assist in getting better pricing service and competitive advantage from vendors.
Besides, dual core design is associated with huge costs. Depending on the underlying topology, the cost of dual core design includes the cost of second planes, operational complexity, sub optimal traffic flow, and convergence.
Dual planes are very useful especially for those situations in which there should not be network failure or in which the cost of usage is not an obstacle. These are complex yet difficult to manage topologies reserved for big companies that data processing is an integral part of their business and for firms that