eBooks

FierceWireless.com 10 9 M AY 2 0 14 M AY 2 0 14 FierceWireless the relationship is to further the adoption of NFV with the help of SDN, according to Pitt. That way, both organizations benefit from sharing technical work and adopting a complementary approach to standards, methodologies and knowledge sharing. NEW WAYS TO DO BUSINESS While SDN may have started as a way to virtualize and improve the operation of data centers, its move into telecom networks is a newer paradigm. "I think networking is becoming part of the world of computing," Pitt said. "The carriers are not used to thinking that way, but the over-the- tops sure do." The ONF board includes representatives from Google, Facebook, Yahoo, Microsoft, Verizon, Deutsche Telekom and NTT Communications and Goldman Sachs; its 11 working groups are tackling issues like northbound interfaces, configuration and management, mobile backhaul and traffic management and traffic steering. "We are solely focused on the commercial success of SDN for the benefit of operators," Pitt said. ONF member company Big Switch Networks is an SDN startup that is partnering with Dell to get its products in front of a broader audience. One of the biggest customer segments for Big Switch is comprised of operators that are deploying LTE because they're finding it's a good time and place to introduce SDN and they need it to monitor traffic, according to CEO Douglas Murray. Through Big Tap, the company offers a network monitoring application for security and performance, among other things. LTE operators in Japan, in particular, have been quick to lead the way on SDN, he said. "The promise of SDN is very bold and it's actually an exciting thing," he said. "Networking has largely been a game of speeds and feeds for the last 10 years. The world has now moved to one where it's less about speeds and feeds, in large part due to virtualization." l Functional and performance testing of network functions For end-users, the primary concerns when it comes to network services are performance of their applications and the quality of experience. They need protocols to work in accordance with specifications, and state machines to be consistent, but are not interested in whether the BNG, routing, CDN or mobility functionalities are implemented in standard servers or purpose-built appliances. For operators there are additional concerns regarding the control- plane and data-plane scale, and whether, for example the number of PPPoE sessions, throughput and forwarding rates, number of MPLS tunnels and routes supported are broadly similar between physical and virtual environments. Testing must ensure that the performance of virtual environments is equivalent to that of the corresponding physical environment. Validating reliability of network services Operators and users accustomed to 99.999 percent availability of physical network services will have the same expectations for virtual environments. It is important to ensure that node, link and service failures are detected within milliseconds and that corrective action is taken promptly without degradation of services. In the event that virtual machines are migrated between servers, it is important to ensure that any loss of packets or services is within acceptable limits set by the relevant SLAs. Ensuring portability of virtual machines (VMs) and stability of NFV environments The ability to load and run virtual functions in a variety of hypervisor and server environments must also be tested. Unlike physical environments, instantiating or deleting VMs can affect the performance of existing VMs as well as services on the server. In accordance with established policies, new VMs should be assigned the appropriate number of compute cores and storage without degrading existing services. It is also critically important to test the virtual environment, including the orchestrator and cloud- management system. Active and passive monitoring of virtual networks In addition to pre-deployment and turn-up testing, it is also important to monitor services and network functions on either an ongoing, passive basis or an as-needed, active basis. Monitoring virtual environments is more complex than their physical equivalents because operators need to tap into either an entire service chain or just a subset of that service chain. For active monitoring, a connection between the monitoring end- points must also be created on an on-demand basis, again without degrading the performance of other functions that are not being monitored in that environment. l Network Function Virtualization (NFV) will enable data center operators and service providers to deploy many virtualized network functions on industry-standard, high-performance servers. In addition to saving both operational and capital expenditures, operators benefit from shortened time to market, the entry of more independent software vendors into the ecosystem, and easier development and deployment of new revenue-generating features. However, replacing thousands of specialized routers and network appliances with NFV-based servers is still a significant challenge for data center operators. Most significantly, virtualized functions, virtual environments and end-to-end networks need to be thoroughly validated prior to deployment. Failures in the field resulting from lack of thorough testing lead to lost revenues and angry customers, and will ultimately slow down the adoption of NFV. Consequently, testing challenges associated with validating NFV environments are not to be underestimated. THE CHALLENGES IN VALIDATING NFV ENVIRONMENT Thorough testing of NFV is essential to minimize network outages post deployment and is critical to its fast adoption. NFV functions and infrastructure must be tested in the lab at the time of service turn-up, as well as constantly monitored for performance. Below are some of the most common challenges and the means of addressing them. Managing the Deployment Challenges of Network NFV Sponsored Content "Networking has largely been a game of speeds and feeds for the last 10 years. The world has now moved to one where it's less about speeds and feeds, in large part due to virtualization." DOUGLAS MURRAY, CEO AT BIG SWITCH NETWORKS

• Cover