September 25, 2018 | Written by: Denis Murphy
Categorized: Cloud Computing
Share this post:
Moving to cloud-based networking with network function virtualization (NFV) and software-defined networking (SDN) demands new ways to enable operations and service lifecycles across hybrid networks. The ability to operate and orchestrate the network through machine learning and artificial intelligence, combined with automation is a key requirement to delivering on the promised business benefits of network virtualization.
To achieve this high level of automation, clients are looking at the DevOps model of software development and operations in which the development and packaging of the VNFs, and design of the end-to-end service chains are performed with runtime operational requirements in mind. Experiences learnt from the DevOps world are now being leveraged into rolling out and managing virtual networking services. Orchestration as part of continuous integration and continuous delivery pipelines; using virtual network function components that can be deployed in smaller units like containers; designing operational requirements into the network service design; and applying cognitive intelligence in monitoring to get ahead of any issues are all areas that need to be considered to achieve that ultimate goal of lights out operations or as some folk like to say zero touch management.
Networking services are multi-tier distributed applications. When you spend time talking to networking engineers, it is clear that they see value attached to network automation tools that are multi-vendor in nature and can be directly tied to the value chain of the end customer. We all know that the NFV orchestrator (NFVO) plays a key role within ETSI MANO, being responsible for the orchestration of NFV infrastructure resources across multiple virtualized infrastructure managers and being responsible for the the lifecycle management of network services. What we also know is that many of the NFV tools today are focused on automating the provisioning aspects, but don’t have much to do with operational state automation. Testimony to that is a recent paper from James Crashaw of Heavy Reading, which states: “As CSPs begin to operationalize NFV, they are finding that some orchestration platforms are failing to manage the entire service lifecycle –specifically, the steps of problem diagnosis, healing, scaling, moving (VNFs from one VM to another) and updating (patching).”
So, let’s look at this world with a new approach.Ask yourself: can we take advantage of the machine to do the automation, as opposed to automating the human scripts or workflows? How about machine-enabled automation of network functions and services lifecycle, driven by a closed loop with cognitive operations to reduce operations complexity and increase speed of innovation? Answering yes to that question will mean network service design, test and deployment are coordinated across service engineering and operations, and you can now bring new services into production with a huge degree of confidence.
IBM Agile Lifecycle Manager is a tool that has proven itself to address this challenge. It provides a services design, testing and automated deployment platform addressing these outlined challenges and complexities of the NFV paradigm. It provides a complete DevOps toolchain that manages the end-to-end lifecycle of virtual network services, from release management of 3rd party VNF software packages right through to the continuous orchestration and running of VNF and Service instances. And at it’s core it has an intent engine that abstracts away complexity by describing how a service should work and allows the intent engine to get you there, crucially without any manual workflow. This results in much reduced operational costs to maintain a service for the duration of its lifetime i.e. lifecycle tasks to upgrade, scale, heal etc..a service.
Machine-enabled automation of virtualized networking when applied, enables faster cycles of new services using less staff and higher resiliency of the network components across multiple domains. As well as the time to value of building a service being decreased, resources required for enablement and operations is also decreased even as complexity of operating environment increases. That is certainly something to think about when choosing a solution.