CenturyLink recently announced a partnership with Ericsson to help revamp their OSS/BSS systems, as described here at Light Reading and here at TMC InfoTech. These announcements are an extension of work that has been going on for quite a while within CenturyLink, and which now will go forward as a part of Ericsson’s Service Innovation Framework. In order to better understand this massive real-world effort, I recently had a discussion with Jack Pugaczewski of CenturyLink, who has had a leading role on this project.

CenturyLink’s internal work and the Service Innovation Framework together form a bold initiative to overhaul the control of network elements and the basic service constructs. They apply a modern, TMN-style layered approach to solving the problem, and they rely on equipment suppliers to provide open Java RMI or RESTful interfaces. I wanted to learn from Pugaczewski how the project is proceeding, and what CenturyLink has learned.

I asked Pugaczewski whether the suppliers have embraced the ideas of open APIs. “Yes, because they understand that this effort is not intended to commoditize their products, but will make them more valuable. Future differentiation will come from software-enabling the network elements for SDN and NFV. Some of the participants have already productized this work.”

CenturyLink’s initiatives and the Service Innovation Framework have received a fair amount of press and executive-level support within CenturyLink, including public statements from James Feger, VP of Network Strategy and Development. How have the rank and file viewed this effort? Pugaczewski noted that “any time there is change, some are excited about it, and some become concerned. The technical barriers are straightforward to overcome. The political challenges are more difficult, especially those related to using automation versus doing something manually. This can create new opportunities.”

One of the points of emphasis in the announcements above was the need for a gradual transition, leveraging the existing systems:

The approach Ericsson is proposing uses the legacy systems as components that contribute specific functions that can become building blocks in a catalog-driven approach to creating new services that it says can speed service activation, reduce complexity, and eliminate touchpoints. Lenahan used TIRKS — the well-known legacy trunk inventory record-keeping system — as an example. It is possible to take the function of TIRKS that identifies available fiber routes between two points of the network and make that a building block in a catalog, to be used to create virtual private network services, for example.

Pugaczewski emphasized this point in his comments. “This project is the automated and layered FCAPS-enablement of the legacy network. The same model is needed for the new SDN / NFV world. Otherwise it is just a science experiment. We have tens of thousands of devices that won’t be replaced overnight. We will have a hybrid of SDN and non-SDN components working together, and physical and virtual components working together. What is needed is a consistent model.”

It is clear that CenturyLink has a real-world focus and wants the Service Innovation Framework to be more than a science experiment, and there has been significant progress towards implementation. Pugaczewski related the following example of the working implementation of the layered approach. “Network equipment connects to the southbound interface at the element management layer of the Service Innovation Framework. At the top will be a business model layer with a northbound RESTful interface to provide network as a service to the higher layers. We can now orchestrate an end-to-end service with cloud components connected to network components.”

I asked Pugaczewski about any issues or surprise benefits that have materialized. He said that it’s been a good experience with few issues. They have achieved a target benefit in “a common service-aware API for Carrier Ethernet using a common XML and XSD. We can now talk in MEF terms and abstract away from the underlying CLI / SNMP interfaces.”

I asked Pugaczewski if he would characterize this project as SDN. “No. It is FCAPS / TMN model network automation that can be applied to SDN or non-SDN components, and to physical or virtual elements. We are going to need network management and service assurance regardless of the nature of the underlying elements.”

So where does SDN come in? Pugaczewski said they deploy “wherever we can, we put in an SDN enabled element as we replace or add equipment. Also, control is not centralized at the element management layer, but rather at the network management layer, where the service request gets broken down and sent to the element management layer interfaces. However, if SDN can make it better, I’m all for it.”

As we wrapped up, Pugaczewski talked about how CenturyLink and the Service Innovation Framework are bringing TMN / NGOSS concepts into the SDN / NFV world. With that evolution, CenturyLink can now build services on top of their four basic constructs (optical, internet, layer 3 VPN and MEF Carrier Ethernet) using modern APIs. Doing so will help CenturyLink bridge the gap between the existing systems and the next generation network they have envisioned.