Operators moving to next generation services will have to support a mix of traffic from the 
base station and in the backhaul. Supporting synchronisation of different traffic types within the same network has created a headache that several proposed solutions have yet to cure. But Transmode’s CTO Sten Nordell tells Keith Dyer that there is a solution available
Keith Dyer:
Sten, can you tell us about Transmode’s role in the mobile space. It may be that some readers don’t really think of you as a supplier to mobile operators.
Sten Nordell:
We have not, perhaps, been that visible in that market and the reason for that is that we have been a fibre only solution provider, and mobile operators have tended to utilise fibre only in the inner parts of their core networks. But a lot of operators have WDM for transporting their traffic, and we are a little better known, and do have a reputation, where our products are delivered for core and aggregation solutions. In fact we have two very knowledgeable mobile operators as customers; Telenor and 3. The access space is new for us, though, and that is being driven by fibre going deeper into that space.
Keith Dyer:
What has been driving the push of fibre closer to the edge of the network?
Sten Nordell:
Quite simply the requirement for more bandwidth as HSPA networks, and coming LTE networks, increase requirements for higher bit rates at lower cost. The T1/E1s most frequently used for backhaul bandwidth cannot scale to meet the traffic demands generated by mobile data and multimedia traffic cost effectively. Adding four or eight E1 lines means that prices will skyrocket, and it is not going to work as a sustainable business model.
And as bit rates rise, and you want eventually to go to Gigabit speeds, then you really run into headspace in microwave solutions. So operators have looked to Ethenet to support those sevices, and the higher bit rates, and that makes fibre far more suitable.
Keith Dyer:
I was under the impression that running fibre to base stations remained expensive, and that there was relatively little of it in the access mix at the moment.
Sten Nordell:
So we are in fact seeing the ‘fibreing’ of base stations. In the USA now 50% of all cell sites have been targeted for fibre. That’s aggressive, but we have certainly heard of operators in the LTE space that say they have very large fibre plans.
Another reason for the increased use of fibre is the sharing of the backhaul pipe amongst operators, to reduce costs. When you can split the cost of that connection thee or four ways, then that also makes fibre much more viable.
So mobile operators, who know they must scale to meet this rapid demand for bandwidth to support mobile broadband 3G and 4G broadband services, are having to off-load this traffic onto more cost effective and scalable Ethernet backhaul. Most operators deploying 3G and planning deployment of LTE plan to do so gradually and will therefore need to support 2G, 3G, LTE, and eventually also 4G, at the same cell site.
This creates challenges in having to efficiently and cost effectively support the varying backhaul requirements of for instance multiple generations of NodeBs and eNodeBs, often for multiple operators at the same cell site.
Keith Dyer:
So although it gives higher bandwidth, moving to fibre doesn’t solve the problem for operators of having to support different traffic types.
Sten Nordell:
Yes, up to now it has been relatively easy for mobile operators. All base stations have used SONET/SDH, whether for fixed or microwave, which is natively built in for handling the handover of voice and data calls. If you turn on a packetisation architecture you have to make choice about how to handle the inherent incompatibility between SDH and Ethernet.
Ethernet to the base station is complicated because Ethernet is not synched in its own right, and is best effort. People have looked at different alternatives, zooming in on Synchronous Ethernet, or Packet Sychronisation techniques such as the IEEE 1588.v2 timing algorithm.
But this is not scaleable as you are faced with two different types of synch, one in old base station sites and one in new sites. That means two different timing mechanisms, one in 2G and other in 2/3G, and the compatibility between the two of them not proven.
For cost effectiveness, operators would like to converge all backhaul traffic from a cell site onto one facility. But not all mobile operators are confident that it is the time to move all their legacy traffic onto an Ethernet backhaul network. So many E1s serving legacy voice backhaul needs are still being left in place today, both to avoid interruption to service and because T1/E1 provides tried and true physical layer timing in which many operators have greater confidence in than some of the newer packet synchronisation techniques such as IEEE 1588v2.
Pseudowire solutions produce an impact on how operators manage, maintain and support their IP networks. They need to deal with jitter, bit error, all the stringent QoS KPIs. And that’s hard for operators to achieve because they do not typically have the skills and technology to deal with it.
Added to this, with cell site sharing, you are not just talking about synch across different traffic from one operator, but supporting multiple operators from the same site – each with their own timing mechanisms. You can see why they would often like to continue to leverage their tried and true T1/E1 physical layer synchronization for voice, at least in the short-term.
Keith Dyer:
So how can operators harness the business benefits of moving forward onto Ethernet, whilst maintaining reliable TDM services?
Sten Nordell:
What we have done is said, why not leave the TDM as it is and natively transport it over Fibre at the same time as Ethenet – side by side over one wavelength. This means you can carry on in the same way you have managed TDM synch, and gradually build out Ethernet to base stations over fibre where it is needed, maintaining synch in TDM-Ethernet environments as you do so. It’s a smarter approach.
The head room that this gives operators means that they are very, very interested in our solution. It means they can utilise that fibre infrastructure more effectively and let them build IP networks slightly differently. Rather than work at the Layer 2 level to all base stations they can concentrate on Layer 1. It also works very nicely for the LTE/SAE architecture.
Keith Dyer:
How does the solution work?
Sten Nordell:
As I said, our Multi-Service Mobile Backhaul Solution has the ability to maintain multiple synchronisation signals to the cell sites in a mobile network – enabling the operator to simultaneously support a mix of 2G (TDM), 3G (ATM/Ethernet) and LTE/4G (Ethernet) radio base station equipment and legacy E1/T1, FE/GE and STM-1/OC-3 for one or more mobile operators.
Our solution achieves this by handling TDM traffic natively – retaining all of the requisite native TDM attributes and physical layer timing delivered by E1/T1s as well as native Ethernet.
Transmode’s solution natively supports both legacy TDM and new Ethernet backhaul traffic over a single wavelength, while supporting up to eight synchronisation streams that can be either physical layer TDM synchronisation, Synchronous Ethernet or packet synchronisation. This simplifies making a gradual migration to 4G and the all-IP network and cost effectively scales to meet the fast growing traffic requirements of mobile data and multimedia traffic while providing operators with simplified OAM&P procedures.
You can think of it as a replacement for pseudowire but it also works with pseudowires, and/or as a transport for pseudowires. It can also complement pseudowire over high-capacity fiber-optic transport.
Another advantage is that it provides fundamental elimination of jitter and bit-error multiplication, thus ensuring a synchronisation signal that meets E1 and T1 standards as well as multiple synchronisation domains, for both E1 and T1 as well as Ethernet. That means that those IP network operational and design issues that I mentioned are taken care of as part of the solution, rather than adding layers of complexity.
Keith Dyer:
Is it an approach that has been tried before – natively transporting traffic types over fibre?
Sten Nordell:
Attributes of it are similar to the way Microwave vendors have approached next generation hybrid access networks – putting Ethernet and E1 side by side in the air. But on fibre we have not seen anyone else try to copy it. This is a first for OEMs integrating these types of hybrid networks. The advantage of mirroring the microwave approach is that operators with fibre and microwave backhaul assets can have a common approach across both transport types.
Keith Dyer:
If operators don’t adopt this approach, what will be the impact on their business?
Sten Nordell:
I have not met an operator who is not thinking about how to solve the issue of moving backhaul from out of a TDM architecture. We’ve seen operators with fixed and mobile assets pull two lines to a base station, one for timing and one for the transport of the data. That’s very costly, but they feel they have no alternative. Throwing microwave bandwidth at it is not the solution because using more capacity in microwave is not cheap. Doing a Gig in the air? That’s a lot, and resources are scarce, with operators facing auctions and increased financial and regulatory decision-making.
We know from a pricing perspective that we’ll prove competitive against other solutions, really because we offer that L1 capability rather than complex L2/L3 solutions. So operators can progress with a much lower power design, compared with competitors on he same cell site, gaining a very good price and power to performance ration. That’s why fibre to the base station is becoming the preferred solution.
Keith Dyer:
And when will the product be available?
Sten Nordell:
We are launching the multi-service backhaul solution officially at the Broadband World Forum, with limited availability by the end of this year. We’ll also be demonstrating the solution at Mobile World Congress in February 2010.
ABOUT STEN NORDELL:
Sten Nordell, Transmode’s Chief Technology Officer, was formerly Chief Convergence Officer and VP Networks & Platform Strategy at mobile operator Telenor Group as well as having run his own telecoms consultancy. As a result, Sten is one of the few people in the industry to have senior level experience both in building mobile networks from the operator’s perspective and in designing systems as a vendor to meet mobile operator needs. Sten has made numerous presentations at conferences and consortia. including the tmforum, and the Ethernet Expo and has also served as a Board member with the IPSphere Forum.