Q&A: TELUS Takes on Transformation

Ibrahim Gedeon is the outspoken chief technology officer for TELUS, the second largest telecommunications provider in Canada. TELUS earns annual revenue of $8.1 billion and boasts roughly 5 million access lines, 5 million wireless subscribers and 1 million Internet customers. Facing increased pressure from cable competition on all fronts, TELUS has been publicized for its aggressive forays into quadruple play services and has been a pioneer of IPTV with partner Alcatel. TELUS is undergoing the kind of transformation everyone in the industry has been talking about. With the help of a number of key vendors, including Syndesis, TELUS is constructing a flexible OSS/BSS architecture using commercial vendor products. The company believes its approach is as future-proof as possible and can support the demands of real-time, application-based services. Gedeon and Syndesis CTO Mark Nicholson lent their time to Billing World and OSS Today to shed some light on what TELUS is doing behind the scenes.

BWOT: TELUS has received a lot of attention for its aggressive service rollouts. What can you tell us about TELUS’ technology vision and what you have accomplished in your OSS/BSS architecture?

Gedeon: Our CEO [Darren Entwistle] decided we needed to be an all-IP network [with the idea that the] Internet experience will be the broadband experience. So we were the first in the world with an all-IP MPLS network for broadband, VoIP and so on. Our CEO also sees our OSS as a differentiator, so we also embarked on a transformation of our IT systems. Our network is IMS-like, and we’ve always had that vision for our architecture. He was clear that we needed a vision to decouple the services from the network and from the user experience. Just like on the network where we’ve transformed and decoupled from the switch and from the transport, we’re doing the same thing in OSS/BSS so that as we swap technologies out we don’t impact the customer experience or what we do in billing.

Some of our peers and competitors have [taken an approach where] they tried to predict the future. Five years later they had what was great five years ago. As a geek I’m glad my CEO is a firm believer in IT, and more importantly in systems—he’ll say “I want an 18-month lead on my competitors in systems,” because that will make or break a company.

We focus on our architecture and our stack, and we don’t have a static model, we have a flexible model. The whole thing was done to loosely couple and go with the natural evolution of technology.

Nicholson: From an OSS perspective, TELUS no longer wants network elements and network technologies to be the products they are offering their customers. In other words, the network is no longer the service but is now one part, an important part mind you, but one part of the service. They are thinking in terms of services and the “friendly user experience” and not in terms of technologies. So the OSS has been built to abstract technologies away from the product catalog and associated order processing. For example, on their consumer broadband offerings they did not want to be tied to any specific form of xDSL technology such as ADSL, ADSL2+, VDSL, etc. even though we were brought in initially to support ADSL2+. In other words, they did not want the access technology to define the solution, but wanted to be able to change that underlying communications technology without changing the notion of the subscriber or the broadband service (e.g., 1.5 Mbps silver service) they subscribe to. In the past, whenever a network rolled out a new technology, many changes were required in the entire stack up to ordering and billing. TELUS has tried to make their OSS stack agnostic and flexible to the underlying technology. When it comes to the network, they can make changes in our application without making broad or sweeping changes to the OSS.

BWOT: You said that in achieving flexibility, one still has to pick certain anchor points that won’t or can’t move out, but that those anchor points must exhibit very specific “DNA” to enable a highly flexible architecture. What “DNA” do you look for?

Gedeon: The DNA is that of a product-based coupling between OSS and BSS; network and applications; and policy and network, all based on loose coupling via SOA. In addition, [the vendors must display a] willingness to work with the standards bodies and a commitment to deploy the standards. From our point of view, TELUS is a services provider that sells telephony and video, rather than a telephone company or broadband company. IMS, or being IMS-like, provides an example of how by using standards we can shift the service provider business from one that is carriage-based to one that is application-based.

BWOT: What are some of your key anchor points?

Gedeon: We have built an overall architecture framework with well-defined vendors in each function. However, we are pragmatic about being right versus being successful. Certain functions and suppliers are sometimes required to chip in out of position. Having a real architecture is important, but we also need to treat each technology, and where it impacts convergence, with consideration to time to market and overall cost of ownership. We’ve identified that we like Amdocs for billing and care. We defined inventory (NetCracker) as a demarc[ation] point, and the data models—legacy or otherwise—are well defined. We have Syndesis doing workflow, service adaptation and activation. Those were clearly defined. We’ve done the work with Syndesis, as had AT&T, and we have Accenture taking the code from Syndesis and making it TELUS-like. Similarly, with Atreus Systems, we were planning on using a vendor EMS for VoIP activation. However, using Atreus’ solution and their portal as an integration back end, we can theoretically change our softswitch vendor without impacting the customer or their experience.

Nicholson: TELUS has brought in Amdocs, Atreus and Accenture to help with the design aspects of a service. We’re the provisioning piece that handles the network. We offer to their other OSS an abstraction of what we are going to activate in the network. Syndesis is doing activation of Cisco Ethernet, ATM, and Frame, Alcatel Ethernet concentrators, and more. It has a wide technical footprint.

We are also handling service adaptation where we are mapping a high-level connectivity request into the access, transport and other related assignments that are needed. If TELUS needs a new technology, we can map it in transparently. We are providing a supporting piece to the workflow and have trained their developers on how to use our IT network and service object models. They can do the workflow they need for things like soft reservations, port assignment, policies and what have you. We have implemented many of these functions in our product, but they can orchestrate them in the way they need to fulfill their specific ordering processes.

BWOT: You also mentioned earlier that you look for vendors who aren’t afraid to bring their expertise to bear when they think you—the customer—made a wrong decision. In what situations has being open to criticism helped you achieve your vision?

Gedeon: If your partner isn’t willing to come to the table and tell you if you’re right or wrong, and if it’s wrong how to make it better, you don’t have much of a partner. We’ve been lucky in having partners that are very up-front. Syndesis helped us see the value of logical inventory versus physical inventory, and abstraction of technology and service; this achieved our time lines, more revenue for them, and it fits within our overall network strategy.

Nicholson: Another thing we debated back and forth was thick versus thin architecture. We normally like to deploy as a thick solution where you give us a macro request for connectivity and we’ll take care of the underlying technology domains. The overall system does not have to deal with detailed, complex assignments such as Ethernet VLANs and service access points, IP QoS and routing attributes, or aggregator traffic marking settings—we can handle that. That offloads a lot of complexity from the other OSS and simplifies what must go into the inventory of record. We can maintain the “service-ready” network.

We pitched that concept to TELUS, but the folks on the IT side, looking at an eTom architecture, thought it wasn’t a good plan. They thought that we should be a thin activator and that they could handle the rest in an external inventory system. They’d already made a major commitment to Netcracker. We said we could do it any way you want, thick or thin. Our architecture let them plug us in either way. And as it turns out, they are using us as a thick model while they roll out things on the NetCracker side. This will be a federation of systems that have the flexibility to move around. If the architects need to make course corrections it won’t cost an enormous amount. This is working, and we’re actually doing it on a fairly reasonable budget.

BWOT: There are many carriers talking about gaining flexibility, decoupling services from technology, and going with a very flexible and loosely coupled architecture. Most of what one hears is spun to the positive, but there have to be challenges for the technology providers. What’s one of the biggest?

Nicholson: I think the biggest hurdle for Syndesis has been change, and the fact that TELUS is aggressive in their realization and execution of their vision. This means that they will refine their course based on market needs, competitive response, lower cost network technology and other factors. They expect their partners to respond accordingly—and I agree with that philosophy.

From a commercial software development point of view, we don’t always have enough detail in the requirements specs such that we can clearly understand what is already in product, what needs to be added to our product (and will be applicable to many other customers), and what is custom and needs to be outside of product. That’s probably the only sort of broad issue in doing an OSS this way. It is perhaps more interactive than what a product-oriented software company would normally want to be. As a product company, it’s a bit antithetical to be developing this stuff really interactively, but we go through this process as we really believe in the product model and the significantly lower TCO our customers benefit from, versus the sometimes greater than 10-to-1 services-to-license cost our competitors bring to their customers.