Service Chaining Replaces Middle Box Chaining in Networks

Share:

By: Larry Loeb |

The actual bit manipulation that a network must do has historically been done by hardware boxes that have specific, predefined functions. Those “middle” boxes — things like routers, switches, load balancers, firewalls and proxy servers — are organized by the network’s architects into functional groups that work at the network’s scale. This is how enterprise networks have traditionally been built over the years.

‘Service Chaining’ to Bring Them Together

Then, according to Network World, the idea came around to chain together the services these boxes were performing. These services might include functions like firewalling, intrusion detection, carrier-grade NAT and deep packet inspection. Each network designer may end up applying the idea of “service chaining” in their own particular way depending on company need, but a few general principles can help make common comparisons.

One of service chaining’s basic assumptions is that it is done on a software-defined WAN (SD-WAN) enabled differently than traditional network. It therefore has the ability to deal with data of varying origin and destinations more flexibly. This logical use of a SD-WAN allows things like network egress points to be defined as a service. Businesses can then form policies in the design that will “chain” these services into the traffic flow for the specific desired application types inside the traffic.

Monitoring real-time path performance can help determine which service is doing the best job for a given task, and identify if new services should be prioritized within the service chain. Techniques like this ensure optimal performance for an otherwise traditional WAN.

Leverage Outside Services

One area slated to benefit from this chaining approach to networks is with an enterprise’s desire to use services that are available outside the network. The goal to replace boxes that may live inside the enterprise’s data centers can allow the company to scale and control services that are otherwise restricted to onsite solutions.

For example, cloud-based solutions for content filtering and access control services are becoming available now, and they generally offer the same kind of performance that was once only available with a physical data center installation. These cloud-based solutions may offer the enterprise a lower cost along with an improved performance for the entire network.

But using these cloud-based solutions on a traditional WAN may be too complex an installation for the IT team to perform and maintain. Not only that, but the specialized tweaking a traditional WAN needs in order to use such a service is usually huge. SD-WAN avoids having to set up and maintain these kinds of specialized tunnels.

Performance Affects Decisions

The SD-WAN network can also use the performance measurements to decide which service to use and automatically handle any fallback scenarios. This is one of the ways the kind of service chaining controlled in the SD-WAN layer is giving rise to some new and sophisticated solutions. Through the use of a trusted advisor, the cost-efficient opportunity in increased network efficiencies can be brought to the forefront when evaluating an enterprise’s overall needs in more depth.

Topics: , , ,