Design, deliver, and run enterprise blockchain workloads quickly and easily.
All servers and systems
Heather Kirksey held up her smartphone. "How often do you stare at your smartphone? How often do you use the Internet on your phone?" asked the vice president of network functions virtualization (NFV) and director at the Open Platform for NFV (OPNFV), speaking at the Open Networking Summit. "That's why you have to care about open source networking. We are transforming the global telecommunications infrastructure."
Perhaps you still think of networking in terms of hardware infrastructure: the Wi-Fi router in your office, the cables hiding in the plenum, or the Internet backbone cable that a backhoe just ruined. However, moving forward, tomorrow's networks will be built from open source software-defined networks (SDNs) running on a wide range of hardware including the open source Open Compute Project (OCP).
"It's exciting to see the principles of open source software development come to hardware," says the Linux Foundation's Arpit Joshipura, general manager of networking. "We see OCP as an integral partner as we explore new opportunities for SDN and NFV deployments."
"Carriers are looking for vanity-free whitebox hardware," adds Kirksey. "We're building a stack for a vertical telecom central office with Open Compute hardware and open source software."
That promised solution enables offices, data centers, and telecom carriers to deliver bandwidth and services. Our ever-growing hunger for network connectivity is driven by cloud computing, the popularity of Internet video services such as Netflix, and the final maturation of 4G and the arrival of 5G, says Chris Wright, Red Hat's chief technology officer. "SDN and NFV are prerequisites for 5G. Without them, carriers can't deliver 5G."
SDN and NFV started with OpenFlow in 2011. OpenFlow was based on a simple idea: to "exploit the fact that most modern Ethernet switches and routers contain flow tables (typically built from TCAMs) that run at line rate to implement firewalls, network address translation, quality of service, and to collect statistics." With that architecture, you could create what they called "programmable networks."
Since then, several open source projects have built on this basic idea of using software, instead of custom hardware, for networking needs. Developers, vendors, and customers are all moving forward with SDN, NFV, and related programs as fast as they can.
Technologists paying close attention to networking already know about SDN—and the major flaw with this rosy vision is that there are well over a dozen NFV and SDN open source projects. Among them: Central Office Re-architected as a Datacenter (CORD), Open Network Automation Platform (ONAP), OpenSwitch Network Operating System, Open vSwitch, and... you get the picture. Some projects cover some of the same ground, such as the SDN programs, Tungsten Fabric (formerly OpenContrail), OpenDaylight (ODL), and Trellis. Even network professionals find it almost impossible to track so many projects, not to mention their myriad acronyms.
In addition, there are two different parent open source SDN/NFV organizations: the Linux Foundation and the Open Networking Foundation (ONF). Initially, the two groups didn't care for each other. When the Linux Foundation's OpenDaylight first appeared, then-ONF executive director Dan Pitt said, "The main difference between our organizations is that ONF is a user-led organization developing fundamental architecture and building blocks, whereas the OpenDaylight project is primarily a vendor-led organization focused exclusively on implementing one instantiation of SDN." When Pitt used the word user, he meant telecom carriers such as AT&T, Deutsche Telekom, and NTT.
What can the industry do to bring sanity, order, and, most important of all, programs that can work together? First, over time, the Linux Foundation and ONF are bridging the gap between them.
For example, Andre Fuetsch, president of the ONF board, and AT&T Labs president and CTO, just announced ONF's new strategic plan at the Linux Foundation's Open Networking Summit in Los Angeles. Phil Robb, Linux Foundation's vice president of networking operations, said at the conference, "Linux Foundation and ONF are both trying to drive innovation across telecom. We're getting into alignment. Our mutual goal is to make SDN work. We want to work together to ensure there's no needless overlaps. We all want to get into production ASAP."
Meanwhile, the Linux Foundation announced a plan to harmonize standard developing organizations, industry groups, and open source projects to capitalize on each other's strengths. Previously, these groups operated autonomously, yet their efforts had to coordinate to create working solutions. In a world that moves at the speed of software, that process was too slow. Close coordination is essential, everyone agreed, and SDN/NFV are too pervasive for any single group to own or drive. While any stand-alone component is important, it can only do so much on its own.
Then, an amazing thing happened: Instead of sinking into the mire of grand technology initiatives, which never garnered mindshare, the Linux Foundation's harmonization plans bloomed.
Kirksey observes, "It was natural evolution to harmonize the governance."
Plus, there are economic advantages for these groups to get outside their own silos.
In January 2018, Linux Foundation and several open source SDN communities came together to form the Linux Foundation Networking Fund. (LFN). It has a two-fold purpose: to encourage SDN groups to cross-pollinate their ideas, and to use a unified administrative structure to save costs.
"Following the example of the Linux Foundation's Cloud Native Computing Foundation, which brought together Kubernetes and other platform-as-a-service (PaaS) cloud projects," Joshipura explains, "LFN will bring similar cohesion to networking communities that in many cases are already working together."
Each project will continue to operate under its own existing meritocratic charters. They maintain their technical independence, community affinities, release roadmaps, and web presence, but staff and financial resources are shared across member projects via a unified governing board.
SDN organizations have been jumping to get on board. Already, nine of the top 10 open source networking projects have joined. In late March, "LFN hit a milestone with 100 members in less than three months," Joshipura says. Among the vendors is ONF, which joined as an associate member. That puts everyone on the same team.
Another benefit is an opportunity to integrate ideas. For example, the ONF's interest in reference architecture in its new Stratum project is being embraced by the Linux Foundation and its groups. As Jonathan Bryce, executive director of the open source OpenStack cloud, says, "Harmonization is mostly about creating reference architectures."
So far, so good, but what's more important than all the project and organizational news is getting SDN communities, management teams, and developers teams together at one place and at one time.
This was the first time we had all the developers co-located, Kirksey says: "Just having all the developers in one room was very valuable. People were saying, 'Oh, we can use that test!' The docs people said, 'We can use the same tooling!'" The result has been a real uptick in development, she adds.
Everyone at the Open Networking Summit agreed. The halls and meeting rooms were packed with people who had spent years working on similar ideas with similar goals and were now finding that their competitors could actually be their best allies.
Of course, not everyone is completely comfortable with this open source approach. Juniper Networks' Tungsten Fabric is a Linux Foundation open source project now, but Randy Bias, Juniper's vice president of cloud technology, told me, "We're committed to working with other projects. However, speaking on my own, there will be competitive projects. ODL and Tungsten's teams must work harder and better in 'coopertition.'"
Still, the true transformation of SDN and NFV is coming not from networking technology nor social networking. It's emerging from that oldest human networking form of all: face-to-face meetings.
This article/content was written by the individual writer identified and does not necessarily reflect the view of Hewlett Packard Enterprise Company.