Operators are moving to LTE and Ethernet backhaul in an effort to reduce their cost per bit, and overall opex costs, but they face challenges to be able to monitor and assure services in the new environment. Jay Stewart, Director Ethernet Service Assurance, JDSU, tells Keith Dyer how they can meet those challenges.
Keith Dyer:
Can you outline what the issues are with service assurance in Ethernet, and how that applies to mobile backhaul?
Jay Stewart:
The pressure to succeed at backhaul has never been greater for service providers – the surge of mobile data traffic and associated cost of its transport has created a number of significant business challenges.
The important, inevitable step of updating a backhaul network from TDM to Ethernet is not about “if” but about “when.” A big problem is that Ethernet was not originally designed for carrier-grade use. Backhaul brings a new dimension to the testing of Ethernet. The introduction of Ethernet to the backhaul adds new complexity, in terms of the physical speeds required and in terms of the new provisioning and set up for the service attributes of that space.
Keith Dyer:
What are the challenges that operators face and how can they overcome them?
Jay Stewart:
Service providers are facing difficult operational challenges with the deployment of Ethernet services for mobile backhaul. The turn-up of 3G and 4G is seeing the offloading of data on to Ethernet, with providers having multiple Ethernet Vitutal Circuits (EVCs) – each one with different Class of Service (CoS) attributes. Those providers are being tasked to monitor that traffic and apply different shaping and management end-to-end across different network elements. That’s been the big change that backhaul has brought to Ethernet testing.
In the past performance monitoring operated at the network level and was relatively simple – but now with next generation elements, performance monitoring is a much bigger issue as operators deal with multiple QoS mechanisms. The concept has moved from being
network based to being service based, which does not equate to what people have used in the TDM world.
LTE for instance very clearly calls out nine QoS classification IDs, and how to manage that is also outlined in 3GPP. So as well as an increase in the amount of data the network has got more complex with more QoS mechanisms, and with everything being IP based.
Keith Dyer:
So that puts operators in a very different place from where they were before. Can they lean at all on their prior methods of assurance?
Jay Stewart:
I think at a high level they can lean on things they have been used to doing – but in going to Ethernet they are necessarily going to have to move to a services based approach to performance management, managing each EVC as a separate link, and managing the multiple QoS mechanisms within each EVC. In each case they must manage each one as if they were a separate SLA.
Operators have found that they have been able to get the architecture to work, but being able to carry out their test and performance monitoring has not been completely thought through. One customer said to me that they have taken 30 years to build their network, and now they are asking me to convert that to Ethernet in three to four years with fewer resources. So they are facing more services over more circuits with fewer resources, meaning a focus on reducing opex and lowering the overall cost per bit.
Keith Dyer:
So operators need solutions that are specifically designed for this Ethernet service environment?
Jay Stewart:
That is right, and for instance our NetComplete Ethernet portfolio, which covers the entire Ethernet services deployment lifecycle, is designed to provide the necessary functions to ensure reliability during the mobile service provider’s Ethernet backhaul transition. These functions enable not only the transition to HSPA, HSPA+, and LTE but also the accurate identification and troubleshooting of service-affecting issues, including Ethernet services turn-up testing and validation combined with service performance and SLA monitoring. By monitoring service performance and SLA conformance a mobile service provider can proactively monitor and manage the service and avoid costly service outages.
Keith Dyer:
Do you have any customer examples of an operator making this transition from a network to service based approach within the Ethernet backhaul network?
Jay Stewart:
I’m afraid client confidentiality means I can’t name names, but JDSU does collaborate closely with its customers for innovative solutions to backhaul. We did recently publicly announce that we were selected by a major mobile communications service provider to support its mobile backhaul transition to Ethernet. So that is one major customer where our NetComplete Ethernet service assurance solution is enabling the quality deployment of bandwidth-intensive mobile services and applications, including mobile digital video and streaming video.
The NetComplete solution also gives the mobile service provider end-to-end automated service turn-up testing and performance monitoring that is standards-compliant with RFC-2544, IEEE 802.1ag and ITU Y.1731, eliminating issues related to network element interoperability and provisioning.
Keith Dyer:
Which takes us into the thorny discussions around OAM standards development…
Jay Stewart:
There has been confusion around the standards. Right now there are standards being developed through the ITU, but they are still not yet being put into network elements, and I think really that is still 12-18 months out. We’re in that grey area of working our how to manage things while the standards gain approval and acceptance within the vendor community.
One problem has been that a lack of standardised OAM tools and procedures, coupled with these more complex IP networks, have led to increasing numbers of technician dispatches for service turn-up, fault finding, troubleshooting, and resolution of service degradations.
Keith Dyer:
Do you feel that there is a danger that mobile backhaul decisions are being made to meet the imminent needs of operators, and that long term planning is taking a back seat?
Jay Stewart:
I think people are living in the now. They know they have got to get bandwidth out to the towers, and they cannot afford to really think about what happens when they move to 4G. LTE brings a big architecture shift, with new interfaces added. But I don’t think it will take away any of the efficiencies we have identified in evolving the performance management environment. We see a roadmap that still needs a lot of work in terms of getting the architecture and services to work – but the main driver is still to reduce the total cost of ownership.
A lot of our work is consultation aimed at that – by bringing an abstracted layer of management that doesn’t have to change as our customers move forward we don’t have to change the way we take data from the network elements. It’s an automated and flexible approach that will stand the transitions that operators need to make.