Automation is all about taking friction and cost out of the business to increase efficiency and improve profitability. As BT's Managing Director and Chief Architect Neil McRae said at Light Reading's Software Defined Operations event in London last November, "Customer expectations are changing, and to meet those expectations, automation is required. Automation is not a choice, it's a freakin' necessity." (See BT's McRae: Vendors Need to Support Automation or They're Out.)
One of the great challenges with automation in the telecom industry is that the networking and IT domains remain heavily siloed in many service providers today with hundreds or even thousands of manual processes required to map data from operation support systems (planning, fulfillment, assurance, etc.) to network management and orchestration systems. Not only does this lead to a lot of "swivel-chair" operations to bridge the gap but fragmented data systems reduce the visibility into real-time service and network state. The quick fix is to over-provision network resources to cope with this lack of visibility but that leads to unnecessarily high capex in addition to the opex overhead associated with highly manual operations.
The longer-term solution is for operators to implement intelligent automation platforms that span both IT and network domains and address key business processes such as plan-to-build, order-to-service, and trouble-to-resolve. The platform should act as an abstraction of the underlying network and IT resources which are presented to various lines of business (mobile, fixed, wholesale, etc.) via APIs to be consumed "as-a-service."
The key building blocks for such an intelligent automation platform include:
1. Orchestration -- an engine that allows operators to design and build workflows which are then pushed to network systems to make configuration, scaling, and other changes, on-demand.
2. Visibility -- an end-to-end view comprising network topology and a federated network/service inventory.
3. Intelligence -- a combination of policy and service assurance underpinned by AI and analytics. This block allows operators to understand what is going on in the network in real time, make predictions and apply appropriate policies. Networks are too complex to be managed with rules tables and scripts. Instead, operators need AI-assisted systems that can capture feedback from assurance systems and determine the appropriate action that the orchestrator must take.
4. Open -- not so much a building block as a design philosophy, the entire automation platform should be based on industry standard APIs (MEF, TMForum, ONAP, etc.) that allow CSPs to easily integrate with their existing systems and make future modifications without egregious change request fees and timelines. Operators should strive to become more self-sufficient and where they do continue to rely on partners should do so as part of a collaborative development.
Automation plays a critical role in the ongoing digital transformation of the telecom industry. IT teams are generally the early adopters, having already embraced cloud-based technologies and agile methodologies to modernize their processes. As network teams catch up with the move to NFV and cloud-native networking there will be a convergence between the two disciplines. As they converge, it is key that the automation platforms operators choose to implement are able to span the OSS and network domains, providing a more holistic and intelligent view of the infrastructure and the services that run on top.
To find out more, view the on-demand version of the webinar Breaking Down Barriers Between IT & Network - Your Path to Digital Transformation.
This blog is sponsored by Ciena.
— James Crawshaw, Senior Analyst, Intelligent Networks and Automation, Heavy Reading