Smartly Designed Networks: How Smaller Carriers Can Benefit from SDN, NFV, and the Cloud

Author: Dr. Vishal Sharma, Principal at Metanoia, Inc.

Metanoia, Inc logo

We show how smaller operators & telcos can benefit from the SDN, NFV, and cloud advancements. Contrary to much popular opinion, we argue that the challenges commonly thought to beset small carriers can, for the most part, be overcome with some innovative thinking, executive will, and strategic action, enabling such operators to emerge stronger and more profitable.


Table of Contents

1 New Paradigms in Networking: Cloudy with a Chance of SDN, NFV!

2 Large Operator Motivations – Flexible Services, Lower Costs, Reduced Deployment Times

3 Opportunity for Smaller Operators – Rural Telcos, WISPs, & Emerging Providers

4 Unique Challenges of Smaller Operators – An Opportunity to Transform!

5 Gazing into the Crystal Ball – A Look to The Future!


1 New Paradigms in Networking: Cloudy with a Chance of SDN, NFV!

It seems that no discussion of modern networking and carrier networks is complete today without a mention of either “cloud”, “virtualization”, “SDN” or “NFV” or some combination thereof, and, in many cases, all of the above! While there certainly has been much industry-hype around these concepts, it is also true that these same ideas have energized the telecom and networking industry in the past couple years, and engendered substantial dialog on how service providers and telcos (and not just data center operators) can also transform their operational models and their delivery of services to enterprises or individual end-users to reap the rewards of efficiency and higher profits. This is great because transformation is the bedrock of a progressive industry (and I certainly like the idea of being part of one!). As the old saying goes “If you’re not moving forward, you’re …” – you get the picture :-).

I will be the first to admit that this transformation will not be immediate, will involve a substantial change in culture (which we shed some light on ahead), and will take some iterations, but it is also true (and many in the industry agree) that it is a transformation needed to move our industry to the next level of evolution – to truly reach that holy grail of being able to offer the “network-as-a-service.” This is the idea that the customer no longer purchases raw connectivity or bandwidth and then either stitches a service together or has the operator do it for them. Rather, the customer now purchases a “network”, on-demand, for specific sets of applications (say, voice, video, critical data, what have you) and all of the underlying connectivity, bandwidth, and quality-of-service (QoS) comes with that “network”.

My goal in this blog post, therefore, is to look first a how some of the larger operators are thinking of, and beginning to apply these concepts to start changing the way in which they build networks and deliver services. We will use this knowledge to focus next on what opportunities there are for a slew of smaller (or, as I like to refer to them, emerging) operators, whether they are rural telcos, wireless ISPs (WISPs) or emerging providers building new networks in various parts of the country (US). At the same time, we will look at some of the unique challenges that these smaller operators face, and reflect on how they could actually overcome these by adopting a more software-centric approach than has been customary for them.

Finally, I’d like to focus on what the future holds for those small carriers (and their executive management) that adopt a bold, forward-looking approach to transformation, transition, and growth.

To start with, it would be useful to level-set by defining a few terms for the purposes of our discussion. Let us start with the term “cloud”, which I define as remotely situated compute and storage resources that can be used by a customer (an enterprise, or, in our case, a carrier). If the resource is owned by the customer or dedicated to the customer, it’s a private cloud. On the other hand, if the resource is part of a shared pool, which serves multiple customers, it’s a public cloud.

“Virtualization” I define as the running of applications and/or services on virtual machines (VMs) or virtual devices (which could be VMs running in a server located in a data center(s)), instead of running them on a device/system dedicated for that application or service.

“SDN” or “Software-Defined Networking” is, at its core, the concept of the separation of the control plane of a network from the forwarding plane, with the control being logically centralized. Here I like the analogy of Manish Singh (ex CTO, Radisys) who calls it the separation of the “brains” of the network (intelligent control) from the “brawns” (the workhorse hardware that moves the bits around).

Finally, “Network Functions Virtualization” (NFV) is the notion of running the software typically embedded in dedicated networking equipment or appliances (that is, network functions, like say, firewalls, load balancing, encryption/decryption, and deep-packet inspection (DPI)) on COTS hardware. Of course, more complex functions such as switching/routing may also be NFV’ed if you will, and can be run on COTS servers, many of which can today handle the needs of real-time communications.

With these formalities out of the way :-), let us take a look at the “why” and “what” behind some of the larger operators’ interest in these technologies.


2 Large Operator Motivations – Flexible Services, Lower Costs, Reduced Deployment Times

As might be expected given the ARPU declines, the rising traffic volumes (thanks to the mobile explosion and our insatiable appetite for video!), and the profitability crunch being experienced by most carriers, a key motivation for operators to adopt these new paradigms is cost reduction – the ability to offer services more efficiently and cheaply than they can at present.

Indeed, researchers and leaders at the forefront of these developments, Dan Pitt of the ONF, Marian Croak of AT&T Labs., and Christos Kolias of Orange Labs., among many others, have, in essence, repeatedly made the following point. That the key motivations for adoption of SDN, NFV, and cloud architectures (the “why”) are:

  • Reduction in the cycle time of deploying services (agility),
  • Ability to compose new and innovative services (flexibility, programmability), and
  • Lower costs (economies gained by the use of commodity hardware)

Given these, it is not surprising that almost all of the larger operators around the world have some form of SDN, NFV, and cloud initiative in the offing, where the goal is to virtualize a part of their infrastructure, move its functions into the cloud, use NFV to run the functions in COTS hardware, and use SDN to control the operation of the underlying network. (Of course, all of this pre-supposes an adequate management layer and the ability to orchestrate the various components, and is the subject of another blog post entirely. :-))

So, several of the larger operators are coming at this from their perspective, as is evident from the following list that illustrates the variety of projects (the “what”) ongoing in this space:

AT&T for example, has had an entity called the Intelligent Routing Service Control Point (IRSCP) for the past several years, which separates the data and control plane to enable distributed control of routing, and can be viewed as an early form of SDN, if you will. In addition, they are now working on an SDN controller that borrows from the OpenDaylight initiative, but modifies its preliminary offerings. AT&T’s goal is to enable what Marian Croak calls the “user-defined network cloud” where they enable their customers to design their own networks and change the properties of that network on-demand.

CenturyLink meanwhile began its foray into this space by first deploying a simple virtual firewall in the cloud (instead of on a routing device), and then built a virtual CDN (content-data network) in the cloud for its internal use for video distribution for its IPTV service. These pilot projects have now given way to larger forays into more customer-facing initiatives that it’s working on.

NTT America started by allowing enterprise customers to activate cloud services under a usage-based billing model. This is different from the commonly used flat pricing model, and provides better overall revenue for the operator and better pricing for customers (who only pay for what they use, when they use it).

Orange, on the other hand, began its efforts in the mobile space, and has what is called a “vEPC” or virtualized Evolved Packet Core (EPC) initiative, where the idea is to take all of the functionality of the LTE EPC (including components like the MME, S-GW, P-GW, PCRF etc.) and virtualize the entire architecture (as opposed to virtualizing each component individually). By virtualizing the architecture as a whole, there is greater coherence in the architecture. Additionally, the customary communications and standard interfaces between the EPC elements can also be optimized.

The setting and needs of these larger operators, with national and/or global scale, are different in many respects from the emerging carriers.

They have large legacy networks in many cases, multiple management and OSS/BSS systems at the back end, several siloed groups managing different aspects of the network (transport vs IP, or wireline vs wireless, optical vs Ethernet), and huge scale, which makes the introduction of these technologies tricky and challenging, not just from a technical perspective, but equally from an organizational and people perspective – changing mindsets, breaking down organizational boundaries, and getting hitherto separate groups to merge and/or cooperate.

The smaller operators share some of these characteristics, but also have some that actually make it easier for them to try these new technologies.


3 Opportunity for Smaller Operators – Rural Telcos, WISPs, & Emerging Providers

When referring to small operators, I will include the rural telcos (defined by the NTCA – The Rural Broadband Association here and here, as almost 900 telcos spread across the US that have on-average 5K lines and gross revenues of the order of $6-10M (although some have revenues substantially above this number, going as high as tens to a hundred+ million dollars), the many WISPs across the nation, and a class of emerging providers that are building new networks across multiple states (e.g LightTower Fiber Networks).

With the ability to now move a greater amount of network intelligence into the cloud, and the availability of COTS hardware in the data center to handle the needs of real-time communication services, it is possible for these operators to think of leveraging commodity hardware to do more complex network functions to cut costs. They can also evolve to being more software-oriented (the “software telco” concept popularized by Metaswitch’s John Lazar, for example), which enables them to add new services with ease, and perhaps think of retiring their legacy central-offices (CO’s).

Of course, these carriers, similar to the big guys, would have to start small, gain experience with virtualization and cloud operation, and then gradually expand their ability to run more and more services off the cloud, leveraging SDN for control and NFV for virtualization.

The following representative examples illustrate areas where smaller operators could start leveraging SDN, NFV, and cloud concepts.

i)                    Intelligent aggregation – this is a concept pioneered by Metaswitch, a well-known vendor supplying rural carriers that has been at the forefront of exhorting rural telcos to eschew their old ways and embrace software-centric network operation. The idea here is that the operator uses a commodity L2 switch, with an OpenFlow interface to a controller, which can intelligently program the switch to offload IP traffic onto the switch, thus moving it away from expensive L3 devices. This is useful to offload video traffic in a local CDN to cheaper L2 devices (as service that many smaller operators offer today as part of their IPTV portfolio).

ii)                  Moving the Central Office to the Cloud – The very “soul” of the rural telco is their central office (as observed by Bernie Arnason in his post here), but with the availability of hosted PBX services in the cloud, it is now possible for forward-thinking operators (not wedded to legacy thinking – a tough one to overcome!) to eliminate their switching gear entirely, by moving it to the cloud on COTS hardware. (This could be in server racks in a part of the operator’s own CO, or a remote data center or public cloud.)

As a result, voice switching becomes a service, the operator can retire old infrastructure, reduce/eliminate maintenance costs, and more towards a more agile voice service deployment. This has the potential to substantially reduce opex and capex (an argument in common with the bigger operators), and allow the telco to concentrate on their relationships with their customers (a strength of being local and embedded in the community), and their physical access assets. There are a number of vendors and professional services companies that can assist telcos with such a transformation.

iii)                Converting the CO to a Data Center (DC): With the availability of space in the CO, due to the transformation above, the operator can now utilize that space for installing server racks and providing data center services themselves to local or regional business customers.

iv)                Deploying SDN-controlled L2 Networks instead of Expensive MPLS Networks: It is possible in some cases, as a local telco PRTC did recently, for even a reasonable-sized telco to bypass the deployment of an MPLS network (which typically has more expensive L3-capable MPLS gear than corresponding L2 equipment), and instead use an SDN-controller-based L2 network to offer enterprise services.

The above representative examples show how operators may monetize the networks already built, transform some of them to reduce operating costs, and invest the savings to fund new initiatives based on SDN, NFV and cloud concepts that are cheaper than traditional alternatives, and, more importantly, future proof.

The question now is, if this is as easy as it sounds, why would every small operator not go this route? The answer lies in understanding how such carriers are positioned, as we do next.


4 Unique Challenges of Smaller Operators – An Opportunity to Transform!

The common refrains advanced in favor of preserving the small operators’ status quo (which, in my view, is equivalent to needlessly ringing their death knell in many cases) is that it is difficult for them to adopt a software-centric approach, that they don’t have access to the requisite skills, that SDN, NFV, and cloud are all complex topics requiring specialized knowledge (which they do, but, as I’ll show momentarily, this need not be a barrier to deployment in small operators), and so on. [BTW, make no mistake, I do realize that in some cases of very remote areas and highly sparse populations, the arguments advanced here may not apply, and, in such cases, support from governmental sources and other supplementary funds may be needed to help sustain smaller, rural operators; but, in a good many cases, a shift of mindset, some innovative thinking, and the approaches outlined here could make a big difference in the operator’s business and performance.]

I propose to argue that all of these are refutable arguments, and that, indeed they should be completely refuted by the executive management at the small operators themselves, if they are to reap the rewards of these technologies to achieve sustained growth and progress, and continue to serve the local communities that many of them have so aptly served for over a 100 years!

Indeed rural carriers have competitive advantages that they can leverage, and access to more resources than they might imagine. In this, Fred Kemmerer, CTO Genband and Bernie Arnason, Chief Publisher of Telecompetitor both agree with my assessment that small operators are a highly resilient and adaptive (you don’t survive a 100+ years without that :-)) set, who are up to this challenge. They first need to think of themselves as “integrated communication and entertainment companies” (and not POTS providers!), and reflect on some of their competitive advantages: that they are embedded in their communities, live and work there, have the benefit of great customer relationships they can leverage, and the advantage, in many cases,  of also selling wireless and IPTV services (making for a larger portfolio and captive customers).

In addition, to the point of being able to bring requisite technical expertise on board, they have several options in this day and age. With the availability of ubiquitous communication (pun intended!), they have access to new technology experts anywhere in the country. Second, with the merging of network and IT concepts in SDN and NFV, it is possible for them to rely more on IT personnel and train them than on traditional network engineers alone, reducing somewhat the burden of hiring “advanced” staff personnel. Finally, there is likely to be a class of technical experts that might actually like to work in the country, if the opportunity presented itself via rual operators willing to transform themselves and hire some bright talent.

Finally, since smaller operators have less complex networks, it is easier, in many cases, for them to experiment and fail fast to discover what works and what does not. Plus, if they so decide, they can be much nimbler than the large operators. Although they may not have the hundreds of millions to invest in experimentation, today they can leverage knowledge available in the industry at large, the move to OpenSource, the free sharing of information in this area, and the availability of industry experts to quickly learn of these areas, determine applicability, adopt what applies, and focus on transforming their networks and services.

So, where many others see “doom-and-gloom” I see  the proverbial pot-of-gold at the end of the rainbow, a chance to be innovative, to be progressive, to sustain, and to remain the anchors of their communities that small operators invariably are.


5 Gazing into the Crystal Ball – A Look toThe Future!

So, is the future of SDN, NFV, and cloud in small operators bleak? Would the small operator become an extinct class? Will rural communities be left at the mercy of the so-called big guys, or without connectivity? Will the small operator networks be obsolete a few years from now?

As I look at my crystal ball, my answer to all of these is “no”. I believe that the small operators are largely resourceful and have shown they can be forward looking (many of them deployed wireless services and IPTV ahead of some of the big guys, in communities that no one would have expected, and many today offer optical, standards-based Ethernet services, which is quite remarkable). And the operators realize the need for change. Such an environment is ripe for them to quickly grasp the value of these new technologies, start experimenting incrementally, reach out to the still developing communities around SDN, NFV and cloud technologies in different parts of the country and in different for a, and partake of these exciting developments to become stronger, more profitable, and more valuable to their communities.


You may also like...

Leave a Reply

Your email address will not be published. Required fields are marked *