SDN and NFV both have architectures that are based on the concept of open software protocols and software applications running on commodity or white box hardware. This architecture provides two advantages to the old network appliance approach of running services on dedicated, special purpose, hardware. First, the transition to commodity or white box hardware can greatly improve problems with non-interoperability between appliances and equipment and second, because of their software based architecture, network applications can be developed, tested, and deployed with high degrees of agility.
What does this mean for the future of data center equipment? There's a widespread sense that specialized gear may be on its way out as ever-evolving software commoditizes the machines of the past. Speaking to Data Center Knowledge a while back, Brocade CEO Lloyd Carney predicted facilities filled with x86 hardware running virtualized applications of all types.
"Ten years from now, in a traditional data center, it will all be x86-based machines," said Carney. "They will be running your business apps in virtual machines, your networking apps in virtual machines, but it will all be x86-based."
Accordingly, a number of open source projects have hit the scene and tried to consolidate the efforts of established telcos, vendors and independent developers to creating flexible community-driven SDN and NFV software. OpenDaylight is a key example.
How OpenDaylight is encouraging commoditization of hardware
To provide some context for Carney's remarks, consider Brocade's own efforts with regard to OpenDaylight, the open source software project designed to accelerate the uptake of SDN and NFV. Brocade is pushing the Brocade SDN Controller, formerly known as Vyatta and based on the OpenDaylight specifications. The company is also supporting the OpenFlow protocol.
The Brocade SDN Controller is part of its efforts to become for OpenDaylight what Red Hat has been to Linux for a long time now. In short: a vendor that continually contributes lots of code to the project, while commercializing it into products that address customer requirements. Of course, there is a balance to be struck here for Brocade and other vendors seeking to bring OpenDaylight-based products to market, between general architectural openness and differentiation, control and experience of particular products. In other words: How open is too open?
Brocade wants to become to OpenDaylight what Red Hat has been to Linux.
On the one hand, automating a stack from a single vendor can be straightforward, since the components are all designed to talk to each other. Cisco's Application Centric Infrastructure largely takes this tack, with a hardware-centric approach to SDN. But this approach can lead to lock-in, and it is resembles the status quo that many service providers have looked to move beyond with the switch to SDN and NFV.
On the other hand, the ideal of seamless automation between various systems from different vendors is complicated by the number of competing standards, as well as by the persistence of legacy and physical infrastructure. Moreover, vendors may be disincentivized from ever seeing this come to pass, since it would reduce their equipment to mere commodities in a software-driven architecture.
Recent OpenDaylight comings and goings
Hence the focus of Brocade et al on new revenue streams and business tactics, which in Brocade's case includes a commitment to "pure" OpenDaylight in its products. The OpenDaylight project itself has had an eventful year, with many of its contributors weighing what they can gain by working on open source standards against what they could potentially lose from the commoditization of the network hardware sector.
VMware and Juniper have dialed down their participation in OpenDaylight in 2015, while AT&T, ClearPath and Nokia have all hopped aboard. For AT&T in particular, OpenDaylight has the advantage of potentially lowering its costs and accelerating its deployment of widespread network virtualization as part of its Domain 2.0 initiative. As we can see, there may be more obvious incentives for carriers themselves than for their vendors to invest in open source projects.
AT&T's John Donovan has cited the importance of OpenDaylight in helping the service provider reach 75 percent network virtualization by 2020. Sustaining the attention and contributions to OpenDaylight and similar projects will be essential to bringing SDN, NFV and cloud automation to networks everywhere in the next decade. DevOps innovation platforms such as QualiSystems CloudShell can ease the transition from legacy to SDN/NFV by turning any collection of infrastructure into an efficient IaaS cloud.
The takeaway: Open source software projects such as OpenDaylight are commoditizing hardware. Rather than rely on proprietary devices with little interoperability, service providers can ideally run intelligent software over standard hardware to make their networks more adaptable. The deepening commitments of Brocade, AT&T and others to OpenDaylight in particular could pave the way for further change in how tomorrow's networks are virtualized.