Open Networking Summit 2016 in Santa Clara, California, was a great show with interesting people and eye-opening talks on the relevance of open source software, open hardware and disaggregation. In short, we're facing a fundamental shift in how we integrate and operate networks. A huge task of operational transformation lies ahead of us. Some presenters referred to this change with the phrase “software eats the network”, which highlights the scale of the shift from purpose-built hardware to software appliances running on standard servers and bare metal switches with a layer of open-source software in between.
Various companies talked about their contributions to the open-source community or about the software architectures required to control, operate and orchestrate networks being based on virtual network functions and utilizing standard hardware.
AT&T's John Donovan outlined the status of network virtualization and its next steps. With ECOMP (Enhanced Control, Orchestration, Management & Policy), AT&T has compiled a comprehensive architecture for operating virtualized networks. John encouraged the community to comment on the white paper which describes it and he indicated the possibility of making it open source subject to industry interest.
Omar Baldonado from Facebook summarized the software and hardware projects which have been completed over the last three years (Wedge, FBOSS) and also highlighted that there’s more to come. He emphasized the relevance those technologies have for FB but also for other companies, as the technologies are open source as part of the Open Compute Project (OCP).
Some aspects of Cisco's open source strategy were outlined by David Warden. He talked about VPP (Vector Packet Processing), which is among the seeding technologies creating the Linux Foundation’s FD.io project. This projects aims at enhancements of IO services in virtualized infrastructures.
These are just some of the companies among many others who talked about their present and future contributions to the open source community. During Q&A sessions some presenters were asked what motivates them to engage with open source and what risks they might face as they lose a competitive advantage by publishing their technologies.
Different answers were provided. Some presenters stated that the benefits from open source in terms of speed and flexibility outweighed the advantages of developing a technology in isolation. There was also the view that companies engaging with open source are still able to differentiate as at least some of the “secret sauce” is kept behind. In addition, open source was seen as a means to strategically impact the direction technologies are taking and as such can enable a company to create a competitive advantage over those companies driving on a dead-end road.
It might be helpful to compare previous strategies aiming at creating a competitive advantage with an open source-centric approach. Years ago, innovation was introduced in a very different way. Communication service providers developed a network architecture and network capabilities which allowed them to compete with differentiated services. They defined feature requirements for network components and made them part of their RFIs and RFQs. Tendering processes, type acceptance and field approval followed before a feature could be used to enable a new services. And competitors required considerable time before they were able to add those network capabilities themselves.
Today a service provider does not try to compete by developing differentiating features in isolation. The service provider differentiates through speed, agility and the ability to introduce new services as fast as possible while also trying to reach a very large customer base. Operational processes are key and any inefficiency will jeopardize the potential of a services factory which can produce new services more or less instantly. The service provider is no longer interested in differentiating features on a network element level but needs solutions which are closely aligned with its operational DNA. This can be achieved by pro-actively contributing to open source with code, such as Telefonica with open sourcing OPEN-MANO, or by publishing network architectures such as AT&T with the ECOMP white paper.
Similar reasoning holds for other players engaging with open source.
In short, innovation, differentiation and competition follow very different paradigms. The value-creation process builds on speed, flexibility and efficiency rather than on feature differentiation. Any stakeholder in our industry will need to realign its strategic engagement and align with the rules of a very different competitive selection process.