One of the key drivers for operators’ virtualization strategies is service agility, and automation is critical. For this to happen, traditional RFP procurement practices won’t fly – operators need to be able to onboard new virtualized functions much faster to enable new capabilities. Right now, this process can take up to 24 weeks. A team of operators and vendors is striving to cut it to a day.
There are a number of key challenges, though, as highlighted at TM Forum Live! Asia recently by a panel of experts — Jenny Huang, OSS/BSS Strategic Standards Convener, AT&T; Frank Massoudian, Principal Architect, Huawei; and Johanne Mayer, Director Product Marketing, Oracle Communications.
Along with other companies, they are working together on an award-winning proof-of-concept Catalyst project to tackle some of the obstacles. The team is looking to drive industry-wide adoption of a metamodel and standardized metadata that will enable same-day onboarding of virtualized functions. They are aiming to automate the entire lifecycle of a virtual function from procurement and onboarding to testing and validation, deployment, configuration and assurance. Initially the focus is on VNFs (virtualized network functions) but ultimately everything including OSS/BSS will be virtualized.
The importance of this work to the industry as a whole is demonstrated by the companies collaborating to move it on: AT&T, Verizon, China Mobile and Orange are already champions of the project, with Vodafone and BT set to join. IBM, Huawei, EnterpriseWeb, Oracle, Amdocs and Cloud Best Practices are all participants and the team is also working closely with other standards bodies, including ETSI, OPNFV and TOSCA.
The Catalyst, titled Enabling the Digital Services Marketplace with Onboarding Automation also won the Biggest Contribution to TM Forum Assets award at TM Forum Live! Asia and the Best in Show award at Innovation InFocus in Dallas, highlighting the cutting-edge and practical nature of the initiative.
Pieces of the puzzle
One of the key findings is that an ecosystem approach is the only way.
As Jenny Huang pointed out, “We knew that the traditional six weeks of provisioning was not agile enough and wouldn’t work but it took us a while to really figure out what the specific problem was.”
She added, “When we talk about multi-vendor solutions, we are really talking about ecosystems. Each stakeholder needs to access a piece of the information at a given point in time.”
Vendors and service providers have typically had a one-on-one relationship or close to it – now the landscape is moving towards more of an Amazon marketplace where operators pick and choose capabilities from different vendors. This means suppliers deliver small pieces of the puzzle, rather than the whole solution. How do you make sure everything works together?
Frank Massoudian said, “We have to change our cultures and organizations so we are working together on essentially a cross-functional DevOps venture because without that, the complexity is such that we can’t really handle it on our own and the operators can’t handle it on their own either.”
Continuous integration and delivery will also disrupt financial models where payment is made incrementally as functionality is being delivered.
Defining the package
As things stand, VNF suppliers don’t have a clear idea of what operators want. The panel highlighted the importance of getting to an agreed definition of requirements.
This Catalyst will clarify what operators require, meaning that VNF suppliers can align and understand exactly what to deliver in terms of issues such as licenses, contracts, metrics and more.
Jenny Huang commented, “The virtual function is almost like the Lego brick. The first step is defining the specification for the Lego brick. Once we know that, a service provider will know how to integrate it into their environment and the vendors will know how to build and package their stuff.”
The team is using TM Forum tools such as the Business Process Framework to create a common language and get more concrete definitions in place.
The bigger picture
The Catalyst also fits into a bigger picture of the work going on around VNF packaging. For example, other groups such as ETSI are looking at functional aspects and the team aims to augment and complement those efforts by addressing the lifecycle overall.
Huang said that the specification needs to be, “A well-defined format that a machine can process and with some critical information in there,” such as the metrics, the management APIs, the SLAs and licensing information, and charging/billing.
The proposed solution brings together standards from TM Forum Frameworx and TOSCA (Topology and Orchestration Specification for Cloud Applications) a data modeling language developed by OASIS for describing a portable application package, along with ETSI NFV concepts. The team is also working closely with open source groups such as OPNFV.
In the video below, Johanne Mayer describes how the next iteration of the Catalyst will demonstrate a VNF package from different suppliers and “how we can basically onboard these and ‘open the package’ – take the information around licensing [for example]and send that to the billing system for settlement purposes and take the testing description and send that to the place where they do the testing.”
She added that the team will show the automation behind that, “Automation is what service providers need to reduce onboarding time.”
Asked whether the outcomes of this Catalyst could drive the whole industry, Huang said, “I do think so because I don’t see any other alternatives here.”
Find out more about TM Forum Catalysts and the latest work in this video series: