An increasing number of telecoms and service providers are adopting what is known as WAN-SDN or carrier SDN. Using SDN, service providers acquire the ability to automate and optimize their existing network and quickly provision new services to their customers. For example, in a multi-service network, SDN can help provision resources to diverse services running on the same IP/MPLS core, based on the requirements of each application or traffic flow.
And yet, despite the advantages SDN provides, SDN controllers have drawbacks. With SDN controllers managing and rolling out network changes such as provisioning of Traffic Engineered tunnels or constraint based paths automatically, network engineers lose visibility into the changes. Moreover, SDN controllers cannot plan the network changes to be rolled out, determine the impact of changes on applications and services or determine root cause if the changes create problems.
This is where SDN analytics help by providing actionable information about both legacy and SDN networks. SDN analytics deliver real-time visibility into both the control plane and data plane metrics such as the current state of the network, network topology, IGP and BGP routes, traffic utilization, latency, jitter and device performance.
Analytics data, which can be collected using an SDN analytics tool such as the Packet Design Explorer SDN Platform, provides network engineers with the management intelligence required to automate new service requests and model planned changes before actually provisioning it in the network via an SDN controller. For example, an SDN controller automatically provisioning a new transport path for a customer can use SDN analytics data to tell which path from a source to destination has the resources to meet the path constraints. And because SDN analytics provides visibility into the current state of the network, network changes can be planned and provisioned only if the change has no adverse impact on the network.
The ability to store analytics data is an additional advantage from SDN analytics. When there is a network issue, network engineers can go back in time and quickly analyze the changes made as well as the impact to find the root cause.
There are several use cases applicable to service providers and telecoms that can be achieved using SDN and SDN analytics, some of which you can read about from our blogs below:
Now that we have covered the need for SDN analytics, the important question is: Where should analytics reside? Should it be on the SDN controllers under vendor control? Or is it advantageous to introduce a new layer for SDN analytics and automation?
Packet Design CTO Cengiz Alaettinoglu talks about this and more in his article, “The Criticality of SDN Analytics Data” published in Pipeline magazine. Read the complete article here:
And don’t forget to share your opinion on SDN analytics with us.