NFV Strategies

Can a Common Core Network Support All Services?

Wireless, wireline, cable, satellite and over-the-top (OTT) providers are increasingly offering similar and often competing services -- data, video, voice and Internet of Things (IoT) services are offered by all regardless of the access method. In addition, extensions coming to wireless standards will actually enable further overlap.

For example, 5G will support high-bandwidth, low-latency and highly reliable services; typically only the purview of wireline providers.

With this convergence, there is also diversity in terms of the value and bandwidth requirements of each service. For example, while voice and SMS have historically been very lucrative in terms of $/GB, low bit-rate IoT is less so.

This begs the question: Can a common core network architecture serve all? The needs are challenging. Let's try to answer that.

Today's wireless core networks are tuned to support mobile broadband subscribers. Moving forward, there is recognition that changes need to be made to serve both new providers (wireline, cable, satellite and OTT providers) and new services (fixed access and IoT). Those changes are being realized at both an architectural level and by point products.

For instance, IMS provides a foundation for new services in a plug-and-play like manner through Application Server (AS) add-ons to a SIP call flow engine. A good example is a wireline telephony AS added to a wireless core network. This plug-and-play approach lends itself well to other convergence types as well. For example, cable operators looking at adding SIP-based voice services to their existing portfolios could consider an IMS/TAS like implementation approach, combined with a video AS.

Service chaining, in its early days, is a way to move from a sequential, rigid services flow to one where potentially high-cost treatments are applied or not, depending on the service, in real-time. Email, for example, could pass through a service chain implementing virus, spam and phishing detection, whereas voice and video could have traffic shaping applied to minimize delay and jitter. In short, service chaining enables cost-effective support of diverse service needs, as is also required when supporting different providers.

The NFV framework introduces cloud computing into the telecom world, bringing with it the inherent capex/opex benefits associated with common platforms and virtualized software, as well as supporting geographically distributed data centers. These are themes well understood by the OTT community. By adopting similar architectural paradigms, one can imagine OTT players being more comfortable in adopting wireless common core platforms for new services such as carrier-grade voice or cellular data. For example, the new 3.5GHz sharing schemes in the US will support new entrants, of which OTTs are candidates, and NFV-based implementations from telecom vendors could be one approach to simplify their entry.

Finally, with the development of 5G standards comes the concept of network-as-a-service (NaaS), or network slices. The fundamentals are quite simple and are based on the idea that each service is different, and thus implementations should be able to vary, to match the cost of providing the service with the value of the service. An MME in today's EPC is well tuned for mobile broadband users but is perhaps unnecessarily sophisticated for a simple IoT service.

So why not support two types of MMEs? NaaS institutionalizes this idea by allowing real-time direction of traffic flows to appropriate nodes in the network based on service identifiers. This may apply well to IoT providers looking at implementing carrier-grade solutions for large-scale rollouts.

Now, coming back to the original question on whether a single common core network can satisfy different types of service providers -- the clear answer is yes. The inherent flexibility provided by IMS, Service chaining, the NFV Framework and networks-as-a-service open up the telecom architecture in a way that is more inclusive of both different providers and different service types.

What's happening in all of these at a more fundamental level is disaggregation, meaning a greater separation of hardware from software, individual functions from each other and an ability to select appropriate paths through a network at a more granular level. This decomposition helps create a more layered network that is inherently more plug and play, and more capable of supporting different use cases for multiple access types. When combined, the result is improved efficiency, scalability and seamlessness, which benefit the provider and, ultimately, the end user.

Going forward, we see consolidation at the architectural and product levels increasing. This means we can envision a common core architecture being more prevalent in the not-too-distant future, supporting IT, telecom and other communications providers' needs. Infrastructure convergence facilitates further services convergence. And the circle is closed.

— Michael Murphy, CTO of North America, Nokia Networks

[email protected] 9/1/2015 | 12:05:26 PM
Difference between Mobile op and Fixed/mobile op TO what extent do you see such trends being shaped by the increasing tendency among operators to want to converage their networks?

Will there be a radical different approach between pure mobile operators and integrated fixed/mobile network operators (converged network up to the very edge)?


And, in theory, which approach offers the greatest operational simplicity?
Sign In