Dynamic networks are increasing, rather than decreasing, the need for federation.
Not only do big bang transformations not work, they can’t work.
A big part of the reason is the inertia of existing systems and legacy systems to accommodate new network technology and services – a harsh reality that has often been left unchallenged (or at least, under-addressed) in the OSS world. Instead, new systems are introduced for each new technology, with a view to transform and move legacy technology to the latest and greatest new system. Yet when that day comes, there is too much bound up in the existing system to be cost-effectively moved to the new system. Instead of transformation, one more siloed solution is – pardon the pun – “ossified” with the need for operational staff to “swivel chair” between them.
Dynamic, programmable, software-defined networks were expected to sweep this problem away, with a mantra that the “network is the inventory”. But few CSPs have the luxury of being able to create an entirely new “greenfield” network and its associated software infrastructure.
Over the last few years, we have seen recognition of this problem, particularly within fulfilment and inventory systems, such as Oracle’s Unified Inventory; leveraging the existing platforms in a federated system with the prospect of retiring underlying platforms over time, as makes business sense. However, in the arena of assurance and real-time networks, this remains unusual.
Real-time Topology and Assurance
Much of the problem here was due to the closed nature of many assurance systems; with access to the data being restricted and with data treated by data owners as sacred and not to be shared. Now, with the increasing need to support analytics and other third-party applications, there is an increased exposure of data on which new assurance solutions can be built. There is also a need for visibility into this data to support on-demand service fulfilment applications and for ultimately closing the loop between assurance and orchestration enabling autonomous networks.
For existing networks, federation offers a route to the fabled “single view”. But this should not be the limit of CSPs’ aspirations. With dynamic networks and services, a real-time model of the network, its topology, and a history of changes are all essential to supporting the goals of greater automation, and greater network autonomy.
The Future is Federated
CSPs require an evolutionary approach to adopting cloud-native networks and their automated operational approach. Federation, the aggregation of existing data and systems, is a key enabler of an evolutionary approach to network transformation and automation. It allows CSPs, with existing multivendor networks and hybrid physical and virtual networks, to gain a single view of the network and forms the basis for Artificial Intelligence (AI)/Machine Learning (ML) on top of the real-world network.
In a new 10-page paper, Appledore explores how a real-time topology of the network, based on data federation can (and should) be an architecture strategy for service fulfilment and assurance. The report highlights how Oracle’s Unified Operations Suite, complete with its recent Federos acquisition, is enabling this. Further, we explain how the federation of real-time network topology is fundamental to the successful use of AI and ML. The paper is free to download, courtesy of Oracle.
Download the paper here.