February 2, 2017 | Written by: Steven Teitzel
Categorized: Network Agility
Share this post:
Last week we talked about reinventing networks and the need to “Commit, Adapt, and Innovate….Incessantly”. Reinventing networks with Cloud-based Networking using NFV/SDN compels you to think across multiple dimensions of the opportunity at the same time. We are seeing three layers of discussion occurring in this reinvention. First, in the infrastructure, much good work through today has been spent moving network functions to virtualized form, to cloud-native, and the issues there. As function providers move from porting functions to being virtualized to cloud native into micro-services, more needs to be done. Another layer of the discussion comes from the often missed topic of engaging with the customer themselves in how best, who and when do we move customers/revenue to these new services (see Craig Farrell’s article on “The dilemma of implementing NFV services: How to speed it up” in TMForum). A third lies between these with how we operationalize all of this increasing complexity of the hybrid network to design and develop, run and operate…ie, the Operations Support Systems or OSS.
Many are writing perceptive insights on the need to update and change OSS with NFV as it quickly becomes seen as the linchpin by some as to why we are slow in moving to cloud-based networking quickly. There are a number of efforts to specify and standardize this and commit to creating open source components to rapidly enable. What we find though is there are overlaps in this discussion, a combining of terms, creating sometimes the confusion that is slowing us down as we pose key areas against the other. Perhaps what we need is to unify the discussion and in many ways unify OSS with common threads and connection points.
As the issue is discussed, three unique but overlapping areas emerge in unifying OSS: (1) managing the service throughout its Lifecycle from ideation to end-of-life, (2) providing operations in a whole new way, and (3) how do we develop the hybrid services so they are ready to be run and operated upon design, ie: Service Life-cycle Management, Agile Network DevOps, and Cognitive Service Operations.
Within each of these areas are key linkages that connect them and yet also blur the lines bringing them together so we can seize the opportunity in cloud-based networking. Service Lifecycle Management is about how to on-board, design, develop, and deploy on a continuous basis in multiple scenarios the network service. Being able to scale a VNF with the VNF Manager is one item but we need to able to do this, to enable active upgrades due to recovery, repair, or upgrade to new versions on a continuous basis. To deliver a service of multiple functions across a hybrid environment of virtual and physical requires a repeatable process, to do this in today’s network requires doing this with automation to drive this continuous orchestration with closed loop automation required at multiple levels of: the infrastructure, the network function, and the service.
In addition, the need to move with agility at scale and bring what is needed for this run time into the development creates a need to leverage what cloud technology has enabled for the world of IT or the ability to do agile DevOps in the network. Tools for catalog-driven service development, on-boarding, deployment are great with common modeling to enable this but a key linkage point within Agile Network DevOps is not just thinking about operations while in design, but security, performance, functional verification integrated into and throughout the development and the deployment. An iterative automated standard way for making sure these elements are not just available but automatically tested as a service is needed so they work as needed by the customer. Being able to do this testing, with agility and at scale in an almost factory-like reusable approach is needed.
Finally operations has a linkage point within that starts with deep analytics but more than for greater insights to act upon. Analytics must also enable foresight to act before it is needed and it needs the power of the machine to find those insights rapidly, seasonally, with the changing behavior of the network. This means the use of cognitive computing to solve for networks like what it does for cancer health services – learning, reasoning, understanding the issues of the network. Cognitive operations for cloud-based networking brings with it a shift to a focus on service management for greater operational optimization but also for higher customer experience. So Cognitive Service Operations brings together this service focus with cognitive analytics in an automated manner to reduce operator involvement and also augment the capabilities of the service operations team.
Here then is another key to the unification of OSS in this work of automation. Automation can be done as trust develops and is experienced but this must be accelerated through tools to simplify as you virtualize. To get cost savings as services are built, deployed and operated they have to be executed repeatedly — automation in testing, automation in closed-loop-automation and automation in cognitive operations. Deep cloud learning shows that as automation increases the speed, then the power, the ability to innovate increases also.
Cloud-based networking thought of in this multi-dimensional manner brings together the thinking across the layers of the infrastructure and the customer together with the service lifecycle. The new platform of a cloud-based network begins the journey needed by service providers but real ‘cloudification’ to enable innovation from the network requires automation and the linkages within Service Lifecycle Management, Agile Network DevOps and Cognitive Service Operations that join together the world of cloud-based networking. Networking in a cloud-based world requires unifying OSS.
As we continue to commit, adapt, and innovate incessantly to reinvent networks we will see the unification of OSS occurring and we will be driven to enable these key linkages across the design and run time environments of the network. Over the next few weeks we will continue with a view of these linkages in this series on Reinventing Networks. We welcome you to see more on these three linkages for cloud-based networking and IBM capabilities in a private lab environment at GSMA MWC in Barcelona to understand how IBM can help you thrive in this new world of cloud-based networking.