Service Assured VCPE: 5 Top Priorities for Enterprise VAS
By Dr. Yuri Gittik, Head of Strategic Innovations, RAD
When network functions virtualization (NFV) and software defined networking (SDN) were first introduced by communications service providers (CSPs), they held the promise of a true revolution for enterprise services nimble networks, fast service rollouts and low-cost COTS servers (white boxes) replacing multiple dedicated appliances–straight out of IT operations. The concept of the white box customer premises equipment (CPE)–that communications service gateway installed by CSPs at enterprise sites or MTU basements–has taken the industry by storm.
"Enterprises stand to benefit a great deal from an accelerated adoption of NFV/vCPE, provided they can rely on their CSPs to follow best practices and close all current gaps"
Five years on and vCPE-enabled business services remain at the top of carriers’ agenda, but some aspects of the new technology have proven to be more challenging to realize than others.
Where’s the problem?
There are some gaps that prevent operators from passing on the benefits of SDN/NFV to enterprises now. Regular white boxes, for example, lack some connectivity functions and access options that were not needed in the IT world, but are most definitely required to deliver carrier-grade business services and cannot always be virtualized. Virtualized network function (VNF) prices are still high and many vendor offerings are in effect, walled gardens.
Another gap is universal access, which in effect bars service providers that have any access infrastructure other than fiber Ethernet, from offering their customers the benefits of ubiquitous vCPE deployment. In addition, there’s the lack of enhanced performance monitoring, diagnostics and legacy TDM service support. There may be solutions to all these issues, but they require multiple, separate boxes (typically from the same vendor to ensure interoperability), each with its own management.
There’s also the question of choosing between centralized and distributed deployments to determine where these virtualized functions reside–at the CSP’s data center or at the actual point of service. Larger enterprise customers typically prefer on-site VAS to control and manage their own applications, to meet IT policy requirements and address security issues.
On the other hand, CSPs prefer to offer SMBs centralized services to benefit from economies of scale and simplicity in duplicating service packages in mass quantities. One size definitely does not fit all and CSPs are struggling to find the right balance to properly address their service mix.
Then there’s the issue of the vCPE business case, primarily due to the high cost of branded VNFs. CSPs are looking at open source VNFs as a possible solution, yet still expect their vendors to provide the necessary support and service.
Following vCPE best practices
Unlike in the world of IT, CSPs are bound by scale, complexity (particularly in the access network that is used to connect businesses to their footprint) and the service expectations of their enterprise end-customers. They, therefore, must ensure carrier-grade NFV/vCPE implementation. This mandates a holistic view that takes into account every phase of the service lifecycle:
1. WAN connectivity: Ubiquitous service look & feel over any access technology and media – passive optical network (PON), Carrier Ethernet, xDSL, LTE, and even TDM access. Such variety of WAN connections and interfaces enables a unified global deployment for any NFV VAS and ensures a seamless service experience for the enterprise, anywhere and anytime.
2. vCPE operating system: Must provide true openness to prevent vendor lock-in, so that the CSP is free to choose whichever element they need – be it a server, VNF or a network orchestrator– to support competitive offering to their customers. It must also ensure high performance of all vCPE aspects, including third-party VNF hosting, yet remain slim and agile.
3. High Availability: To ensure maximum service uptime and reliability, covering such aspects as backup and redundancy of the network, connections, vCPE system uptime, NFV infrastructure (NFVI) stability, VNF performance, including service-chained VNFs, and resource utilization in real time. Other required capabilities include diagnostics, troubleshooting and self-healing/recovery (SH/SR) on every level: vCPE, NFVI and individual VNFs.
4. Security: A range of security measures, from TPM to secure tunneling/ VPN and management channels over public networks to allow direct and secure connection to data centers.
5. Automation:To capitalize on the agile nature of NFV/vCPE and roll out competitive services within hours, CSPs must be able to practice zero-touch provisioning, VNF onboarding, instantiation and chaining, as well as maintenance, updates, rollback/ reconfiguration, and tear down. Most enterprises also expect to manage the addition and removal of service elements by themselves, via a self-service portal.
So, where do we go from here?
The vast majority of service providers intend to deploy vCPE-based VAS within the next few years, as it allows them to offer competitive services to support their enterprise customers’ business goals, who can also exert greater control over the service via improved digital experience. Enterprises stand to benefit a great deal from an accelerated adoption of NFV/ vCPE, provided they can rely on their CSPs to follow best practices and close all current gaps.