IMS is more of a guideline to the future than a specific battle plan for taking on the world’s telecom infrastructure overnight. Certainly service providers have bought into it, so their decisions continue to be guided by it, but then they also have differing interpretations of what IMS is and what stage it’s currently at. IMS is coming about very slowly. Legacy equipment is not going away at lightning speed. And how does one market an abstract, worldwide service architecture, anyway? Obviously, you market it to service providers because they can test and prove out a service quickly as a result of having an infrastructure based on IMS. It ultimately benefits users, even though the first apps ported to IMS will be existing revenue-generating applications such as voice (duh). But before we get to that stage, interoperability testing must still be done and interworking/transcoding schemes worked out. Even so, IMS certainly is promising.

 

Tekelec (News - Alert) (www.tekelec.com) is a network applications company with experience in session control that’s helping service providers transition to IMS networks. Travis Russell, Tekelec’s Manager, Product Marketing, says, “IMS is really an implementation standard for VoIP in a wireless network. It actually doesn’t bring any new technology to the table; if one reads the 3GPP specifications for IMS, it calls out many existing protocols used in VoIP today, with new requirements for routing, authentication and authorization. The intent is to standardize how SIP gets implemented to ensure that wireless networks remain secure and reliable.”

 

“While it’s true that IMS enables many new services, the same can be said about conventional IP-based networks today,” says Russell. “In fact, many operators counter that they can deliver everything that IMS promises to enable without deploying an IMS architecture. This is true, although the problems with this approach are many. We already know from our experiences with the Internet that ensuring services are delivered only to those having authorized access is troublesome.”

 

Russell elaborates: “Conventional VoIP works well in a network where subscribers are not paying for each service. For example, the ‘all you can eat’ model where subscribers pay a flat fee for access to various services. However, in our telecom model where revenues are generated through service usage, new mechanisms are needed to prevent fraudulent access to these services.”

 

“The Call Session Control Function [CSCF] introduced in the IMS specifications provides a means for operators to better manage all of the traffic in their networks,” says Russell, “and collect the revenues for that traffic reliably. It also ensures that the operator doesn’t have to purchase multiple OSS/BSS to handle the various functions such as provisioning and billing based on technology. In the IMS model, everything is shared across all services. This means that one billing system is capable of supporting voice, video, e-mail, messaging, and any other service the operator is providing to their subscriber base. This is not the case in today’s model where each service type brings its own proprietary requirements for OSS/BSS.”

 

“There are several new options outlined by 3GPP for charging of services in the IMS model that solve the age-old problem operators see today with multiple billing platforms,” says Russell. “By standardizing on the way the various network entities report and generate billing records, operators can deploy one charging platform for the entire network. This eliminates the need for multiple rating, mediation, and billing systems in the network.”

 

 

At Oracle (www.oracle.com), Indu Kodukula Vice President of Product Management for the Oracle Service Delivery Platform, says, “Initially, when IMS was formulated, everybody thought it was going to be the solution to all the problems the telecom industry had about rolling out next-gen data services in a portable manner. What we’re seeing is that the principles of IMS, with regard to the separation of control between the service layer and the full network, are very much in evidence. We see that across the board.”

 

“IMS rollouts literally follow in the specifications laid down by 3GPP and 3GPP2. From our perspective, we see three main reasons for that,” says Kodukula. “First is that the rollouts have been pretty expensive. Interoperability has been a challenge and even devices haven’t really been available that are IMS-enabled. So rather than rolling out a full-fledged IMS network that could easily cost hundreds of millions of dollars, many providers are taking a wait-and-see approach and they’re trying to find ways they can roll out an ‘IMS Lite’ network or something similar where they can test out the feasibility of IMS-like services without the expense of IMS.”

 

“Second, it’s not that clear what will be the compelling or ‘killer’ IMS applications,” says Kodukula. “Rather than a roll out costing hundreds of millions of dollars and investing that kind of capital expense, what we’re seeing are operators adopting a more horizontal platform which is a service delivery platform that really provides network agnosticity and supports portability of services so that when IMS ‘happens’, they will be able to migrate services using SDP [Service Delivery Platform] technology on top of the IMS network without having to rewrite the applications.”

 

“The third dimension that we see to this is that no one is switching to an IMS network overnight,” says Kodukula. “Initially some providers and vendors in this industry wanted to do a ‘forklift’ of their existing network and essentially plop in a new, IMS network. Definitely in the last 12 months, however, the trend or mood has changed. Now operators increasingly want a way to build and deploy IMS services on the network without having to do a forklift, thus taking an elemental, step-by-step approach to IMS rollout.”

 

“Still, we feel very positive about this market,” says Kodukula. “We provide a standards-based Oracle Service Delivery Platform as a way to build portable applications that run on today’s network and that can also be delivered on next-gen networks such as IMS, SIP, WiFi and WiMAX. We offer a flexible platform that enables portability of services across multiple networks. That’s something we see resonating very well with our service provider customers.”

 

Dialogic (News - Alert) Corporation (www.dialogic.com) makes open systems platforms that enable converged communications so that service providers, developers and system integrators can deliver services content and applications using multimedia processing and signaling technologies.

 

Alex Mushkin, Dialogic’s Product Manager for Multimedia Platforms, says, “IMS definitely has useful aspects. As I see it, there are two sides to the matter. There is a technical side that makes IMS make sense, which involves the infrastructure, in particular wireless carriers moving to the IP-only infrastructure. And there are cost savings after you move everything to IP, which I think is pretty much proven at this point. On that side, those drivers are definitely making things happen.:

 

“On the other hand,” says Mushkin, “on the business side, there’s a bit of uncertainly and ambivalence. Carriers are pushing for an open standard that presumably gives more freedom, and yet they’re still trying to maintain a universal solution within a sort of virtual walled garden. That’s an obstacle on the business side.”

 

“The bottom line is what consumers are willing to pay, says Mushkin, “and right now they don’t seem to be willing to pay a great deal. But there are some areas that we see may be useful and those are telephony-related, such as call centers and PBX features, that are not easily replicated in a pure IP environment.”

 

“Security and billing were challenging in the old days of telecom and they will probably continue to be until they problems they pose are completely solved, be it for IMS or any other specification,” says Mushkin.

 

“Obviously, the whole infrastructure won’t be completely replaced in a year, five years or even ten years,” says Mushkin. “It will be a cautious, gradual process. But on the road to that day, our customers are discovering that even just pieces of IMS deployed can be useful. The media server is one such useful element. Some billing-related features are another, as are related features and interfaces, such as what one sees with the DIAMETER protocol. IPv6 is also an example of a useful feature that can evolve even if IMS doesn’t materialize in the way it was originally designed.”

 

 

 

Centillium (www.centillium.com) is a communication integrated circuit (IC) company that focuses on developing technological solutions that expand communication bandwidth to the Internet.

 

Centillium’s Director of Technical and Strategic Marketing, Dr. Majid Foodeei, says, “We take a unique semiconductor, ‘chip’ view of all this. I’ve been active in the VoIP segment, where many IMS-associated things have already happened -- for example, moving to an all-IP environment, or smarter customer premise equipment, or the role of SIP and other common technical mechanisms. Our top three customers in the VoIP area include our two top OEMs, Alcatel Lucent and Ericsson. So as a chip vendor we have seen a drive toward IMS. Our chips often end up in media gateways, which serve the ‘multiple silos’ of fixed and mobile services. There’s a whole host of access types. There is a big drive among these OEMs to merge their platforms. So, when you come across mergers, such as Alcatel Lucent and Siemens Nokia (News - Alert), they had departments of fixed communications, departments of wireless — probably multiple flavors of wireless, from CDMA, UMTS and so forth, and slowly the drive from the organization’s top is that they’ve got to have a converged platform.”

 

“Service providers are making their decisions under the guns of layoffs and cost reduction, and driven by the accepted guidelines of a gradual adoption of IMS and IMS-type functionality that service providers go by and have accepted over the longer term,” says Foodeei, “they really make their decisions based on those. Having said that, there is a contradictory thing that we see, which really slows this process: all of the legacy deployments for fixed and wireless, which have tons of customers. So doing a converged platform doesn’t happen overnight.”

 

Foodeei goes on, “Nevertheless, Centillium and its partners have developed for our Entropia System-on-Chip [SoC] solution, IMS-on-a-chip single common software for multiple and converged applications that includes various feature sets for totally different applications — believe me, the VoIP environment nowadays is a monster in terms of features, especially when IMS is involved, and you look at the different kinds of cross-connects. [Note: IMS-on-a-chip software encompasses common VoIP for voice/fax/modem/text services, NGN-VoIP, mobile GSM/3GPP, mobile CDMA2000, and an Advanced Media Resource Function/Server. The Entropia IMS-on-a-chip software includes multi-application DSP and network protocol features and provides cross-connects and network interworking and transcoding, such as IP-IP, IP-ATM, and TDM-IP/ATM, with high voice/service quality and low delay.] There are about 700 call connection models involving ATM, IP and TDM, involving a CALEA leg, and handover, different codecs and protocols activated because Megaco/H.248 or SIP can send any message at any time. We had to do innovative things at the software level to make a platform enabling the media gateway guys when they really want to do it and some of them — such as the Alcatel Lucent platforms that we’re in — are certainly capable of all these services, and that becomes the foundation for at least one element of IMS, which is the Media Gateway and the Media Resource Function for Media Services.”

 

“We see IMS slowly become part of the reality of products,” says Foodeei, “Also, customers of our customers will see the benefits of this, because when one buys a platform it can be reconfigured and be used for multiple applications, with a centralization of the solution.”

 

 

Leonid Burakovsky, Director of Mobility and Convergence at Juniper Networks (News - Alert) (www.juniper.net), says, “To be precise about the terminology, we’re talking here at Juniper about Fixed-Mobile Convergence (FMC). IMS is one of the important pieces of this, but not all of it. Now, IMS is ETSI TISPAN-focused [Note: ETSI TISPAN defines the core competence for fixed networks and for migration from switched circuit networks to packet-based networks with an architecture that can serve in both.] More and more we’re see that ETSI TISPAN has adopted IMS as a major single unifying architecture around the notion of ‘service’ and session control. In the old days of Class 4 and 5 switches you had the notion of call control, now you have the notion of session control. SIP is the transport control protocol that conveys this functionality between IMS entities. So you can correlate into the same session you can connect voice, video and whatever multimedia, all of them to the same session.”

 

“It’s important, but from our perspective, it’s only one of the pieces of FMC,” says Burakovsky. “First of all FMC involves a common IP transport layer. This is extremely important and it has been adopted by all the standards bodies. Now, there’s the matter of how you logically separate the IP transport layer from your control signaling layer. This is where ETSI TISPAN did a great job in introducing policy decision and resource admission functionality. This didn’t exist in IMS in 3GPP Release 4 or 5, but the latest releases introduce this functionality. What this means, essentially, is that you can run whatever control system you would like on top of it. It can be just a TDM emulation; after all, TDM is not going away soon. Or it can be a softswitch and a lot of people are arguing whether a softswitch is IMS or non-IMS. But it doesn’t matter, really, because in my understanding of this, IMS concerns the notion of a ‘session’. It should be SIP-based. If it’s SIP-based, it’s IMS. If it’s not, such as SS7 over SIGTRAN then it’s Release 4. But it doesn’t matter. It belongs to the same FMC environment. Nobody is going to throw away the softswitches. So you can have TDM emulation, softswitches, pure IMS systems, or a service-oriented architecture that’s not SIP-based but one that will run and even grow.”

 

“Or take IPTV (News - Alert) systems,” says Burakovsky. “These are not inherently SIP-based but they can use exactly the same policy decision and resource admission functionality and exactly the same IP transport layer. So, our perspective is that ETSI TISPAN was exactly right in defining this whole different control system and IMS is very important, but everything can run on top of the same logically centralized policy rules — although perhaps physically taking the form of several separate devices — and decision functionality and resource admission logic controlled by the same policies.”

 

“On a logical level, what’s happening here is constraining the language of application and services and, of course applications and service demand from the network resources,” says Burakovsky. “When obstructing this language, from the language of network resource enforcement, how are you going to enforce this demand from the fixed-mobile and the cable industry, because cable is progressing in the same direction in PacketCable 2.0., which is really very different from PacketCable Multimedia? So there’s also a fixed-mobile cable convergence [FMCC] going on now too.”

 

“We are looking at the real convergence, FMC and FMCC, and what the logical functionality will be wherein people can really converge different controls systems along with policy decisions and resource admission,” says Burakovsky. “IMS being SIP-based is extremely important, focusing on services, but it’s only one of several possible control systems. That’s in addition to IPTV, Service-Oriented Architecture [SoA], whatever multimedia people will run, and today’s softswitches. And more and more, softswitches are using exactly the same functional architecture that you see defined for FMC/FMCC/IMS stuff. During the transitional era that we are in, softswitches can reuse exactly the same policy decisions and resource admission and exactly the same IP transport layer.”

 

“From this perspective, we do see that people are consolidating on this concept, from all industries: from fixed-mobile and cable, even though the progression may be slow,” says Burakovsky. “But from another perspective, a great deal is happening; for example in the femtocell area. In the beginning people thought, ‘Okay, let’s connect femotcells to the current RNC [Radio Network Controllers].’ Then they said, ‘Oh, we have some technical problems. RNCs can’t scale because they were designed for a completely different architecture, which didn’t take into account millions of subscribers’. The next approach was UMA [Unlicensed Mobile Access Architecture]. Does it work? Absolutely. It’s a good technology, but more and more operators want to connect directly to IMS, or to get VCC [Voice Call Continuity] as a phone moves between cellular and WiFi coverage. So we see that, sometimes, even things that were not discussed a year ago now are a high priority, such as femtocells. But as a network concept, I think this is an absolutely correct concept. Sooner or later service providers will be there and the most important right now is the transition part; how they can transition the networks today with the common goal of FMC.”

 

 

Sprient Communications (www.spirent.com) is a global provider of performance analysis and service management solutions. Andy Huckridge, their Director of NGN Solutions (and columnist for this magazine), says, “Both vendors and service providers must move to an architecture where they can offer services, since services are seen as revenue. These vendors and service providers compete with the Web 2.0 companies, so if they don’t move to an architecture that basically gives them the ability to more efficiently offer services, then they’re out of business. Moreover, FMC is why we’ve seen this, and it has emerged as one of the first ‘services’ that’s really an architecture in its own right. With the appearance of FMC, service providers want to be able to offer complementary services across both their networks – if they have different networks – and if they don’t, to be able to compete with the other types of service providers such as wireless versus wireline.”

 

“So that’s the five-second summary of why all of this has to take place,’ says Hudkridge. “But there’s another argument that’s interesting. What I’m referring to is that as everyone drives forward with new technologies, and we don’t have either the people or the specialists of the old technologies anymore. Sometimes network equipment vendors will say to service providers, ‘We can’t support you on the systems that you bought 20 years ago. It’s time to move forward’. These situations aren’t in the public eye as much, but they definitely occur within industry bodies, forums, and especially the testing industry.”

 

“It’s a real ecosystem of companies helping and leveraging other companies. My company, Spirent, fits in there as well. We’re a next-gen testing company. It’s important for us to provide new solutions to test these next-gen networks and services.”

 

Thus, IMS appears to have a business case, though not quite the “lightning strikes a goldmine” scenario desired by some of its more fervent enthusiasts.

 

-----