Summary Bullets:
- Transport SDN promises much shorter-term rewards than broader high-profile efforts.
- Progress towards transport SDN can be seen in the demand for agile transport services, standards work and deployments this year.
Over the past year, network systems vendors have touted high-profile strategies to transform the service provider business model through extending software-defined networking (SDN) beyond the data center to support network functions virtualization (NFV) throughout the wide area network (WAN). The Tier 1 service providers appear to be embracing these same strategies as they seek to leverage virtual network functions (VNFs) to create new differentiated services. However, these broad efforts can be extremely complex, will likely require fundamental changes in the service provider business operations and appear to rely on a seemingly endless set of standards to avoid vendor lock-in. No wonder service provider SDN and NFV seem to be so far off.
However, another SDN trend has been taking place at the same time, one that promises much shorter-term rewards. It appears the transport SDN is still SDN’s low-hanging fruit and is of immediate interest to service providers. This progress towards transport SDN can be measured in three ways.
Near-Term Demand: At the October 2014 SDN & OpenFlow World Congress, one of the panels addressed the possible first concrete steps towards SDN in the short term. The service providers on the panel maintained that bandwidth on demand (BoD) and multi-layer network control and optimization were the primary use cases for SDN at this time. Transport SDN is the simplest method of supporting these applications. A similar application that was announced in late September 2014, and addressed at length at the November 2014 MEF GEN14 conference, is the Third Network, the concept of providing the ubiquitous on-demand connectivity of the Internet with the assured performance of the MEF’s CE 2.0. Though the full concept of the Third Network is broader than simply “Carrier Ethernet on demand,” the real-time control of Ethernet connections would be a logical first step towards the Third Network. Once again, this use case can be satisfied by transport SDN.
Standards: From the initial considerations of SDN for service provider networks, the industry has recognized that the development and acceptance of SDN standards would be a challenge. Furthermore, the more complex the scenario in which SDN is used, the more numerous the required standards, which threatens to delay any implementation of the technology. Even if the application were relatively simple, such as BoD, a standard for interconnecting SDN domains across multiple operators would be necessary. To address an inter-domain interface standard, the Optical Internetworking Forum (OIF) and the Open Network Foundation (ONF) partnered to conduct the Global Transport SDN Interoperability Demonstration, which defined interoperable interfaces (based on prior work of the ONF Optical Transport Working Group) that could form the basis of a standard for application-based BoD between data center sites (cloud bursting). The success of the testing indicates that the industry could be close to establishing an interface standard.
Deployments: Perhaps the ultimate measure of progress towards acceptance of a technology is its deployment in live networks. In the case of transport SDN, Japanese service provider KVH announced that it would employ a Cyan SDN solution to provide “etherXEN” burstable Ethernet in August 2014, and Huawei and China Telecom announced, in November 2014, the deployment of transport SDN offering BoD and IP/optical synergy (multi-layer orchestration). In the coming months, we expect more Ethernet-only service providers (because of the perceived simplicity) to announce Ethernet BoD using transport SDN.
More elaborate implementations of SDN/NFV are likely to grab the headlines in the coming year, but transport SDN will probably make the most direct market impact. After all, it’s the low-hanging fruit.