Global IP Peering - Bridging Islands Of IP

15 November 2010 |


The rise of VoIP and heterogeneous networks has resulted in islands of IP that present a battleground for those enabling or regulating carrier interconnect

There is mounting evidence that the transition to IP is reaching critical mass. Dell’Oro Group reports that legacy TDM traffic in the wireline voice core market will be more actively pushed out of networks as post-recession investment shifts towards services and signalling based on IP, SIP and IMS.

Telegeography supports this view, with growth in international long distance (ILD) voice minutes having slowed to a crawl, although there has yet to be a decline. “Our estimations suggest that this may happen some time in 2011,” says Telegeography research director Stephan Beckert. “But this is just an informed guess.” Some 25% to 30% of ILD traffic – about 110 billion minutes – is currently transported as VoIP, and while the volume of traffic routed over IP links is growing faster, Beckert says that placing a timescale on the shift is difficult. “As some carriers have large market shares, the transition to VoIP will take place in several ‘lumpy’ steps,” he says. “I suspect that there will be a point when many of the larger carriers convert fully to IP equipment. However, when that will be within the next one to four years is difficult to tell.”

Meanwhile, Telegeography estimates that international internet bandwidth grew 55% in aggregate between mid-2009 and mid-2010, an increase of 13.2Tbps. Private data networks – networks operated by corporations, governments, universities and large content providers such as Google, Yahoo or Microsoft – accounted for about 19% of international network capacity. 

Mobile sends right signals 

According to Dell’Oro Group, the requirement for interconnection via a model such as IPX will be driven by the wireless core voice market. “On the mobile side where you have the subscriber growth, you will see a large-scale move to IMS as mobile network operators look to collapse their networks. And IPX peering from a mobile perspective alone will become critical with the arrival of voice over LTE (VoLTE),” says Greg Collins, VP at Dell’Oro Group. “I think operators may begin to experiment with it early on, but volumes will only occur when you have VoLTE handsets. And that will take a few years yet.”

In the US, AT&T has said it will spend $700 million this year in preparation for its LTE deployment in 2011. And with the integration of LTE into its IMS architecture already underway, it plans to introduce VoLTE by 2013. Rival carrier Verizon is also reported to be planning for VoLTE, although voice calls will run over its existing 3G network for the time being. However, the benefits of IPX are already being pushed today by players such as Sybase 365, which claims that ultimately IPX will offer mobile operators lower costs and easier management of roaming services delivered via a single connection to multiple sources.

“IPX doesn’t do anything that you cannot already do,” says Bill Dudley, group director, product management, at Sybase 365. “But think of it this way: throughout this decade, IPX will become the preferred infrastructure to enable secure and managed connectivity between network operators and application, enterprise and cloud-based service providers. Applications in the IPX connected ecosystem will encompass voice transport, signalling, messaging, enterprise and cloud connectivity, mobile commerce, video and more.”

Dudley describes IPX as a “separate internet” interconnecting smaller clouds of 3G and Wimax wireless networks and FTTx, and points out that it was built to create business, and to transport data between businesses in a completely managed and secure environment not directly addressable by the public internet. In the mobile sphere, he believes an increasing amount of signalling will be handled via the IPX or similar platforms. “Much of the mobile roaming signalling is moving off SS7/C7 and onto IP or IPX, because we have a secure channel with guaranteed QoS between the endpoints. And even the voice call can be transported over IPX back to the home network and originating caller. These might be basic services, but they can use IPX.” 

Solving the cross-network challenge 

Many commentators argue that IPX or other cross-network IP-based interconnect models will be just as important in fixed-line NLD/ILD transport. With telephony increasingly software driven and applications being decoupled from infrastructure, the resulting lower cost of entry has prompted a rapid rise in the number of application service providers who offer a combination of voice over IP, Web 2.0 and over-the-top (OTT) services.

“The number of VoIP islands is increasing,” says Diane Myers, directing analyst, VoIP and IMS traffic, at Infonetics Research. “Traffic is growing up as IP, but some carriers only have TDM interconnects, which means a call originated on a VoIP line could be translated to TDM for ILD transport, then back to IP for termination onto a VoIP line.” According to Myers, current VoIP networks are high quality, but each TDM translation has the potential for something to go wrong. “If the carrier can keep it all-IP and not have to do those conversions end-to-end, then a better quality call is delivered, whilst the carrier can reduce both the cost of termination to a TDM line and the amount of equipment required within the network to conduct TDM translation.”

Myers believes the explosion in text messaging volumes will drive a migration to IP for transport. This trend is already in evidence at BICS. “GRX and MMX use an IP interconnect, so MNOs are now targeting their signalling businesses and their voice businesses,” confirms Dimitri Repinec, product manager for mobile data at BICS. However, he stresses that IPX is much more than simply a hub. “IPX is not voice hubbing; it is an IP cloud where MNOs can implement bilateral models. Next to that we have an IP environment with various hubbing services that are IP-based, but which have a different interconnect scheme – what we call ‘one-to-many’ (hubbing) – whereas IPX is many-to-many.” According to Repinec, the many-to-many scheme (ie multilateral) has yet to gain traction because MNOs have yet to introduce the services calling for such a platform. 

Level the playing field 

The reticence of many service providers in respect of the multilateral approach is one that is raising concerns. “The number of voice service providers is multiplying rapidly and that makes it very difficult for anybody except the biggest players to negotiate and arrange interconnection agreements with all the other players,” says David James, principal analyst, wholesale telecoms, at Ovum. “To be practical, you need a hub-based solution, whether it is IPX or otherwise. But I don’t see everything going to one or the other; it will be a mixed economy.”

According to James, there are other companies now offering vendor-specific or software-dependent interconnect solutions that provide feasible and flexible alternatives to support the multiplicity of different services available. Houman Modarres, director of product marketing at the IP division of Alcatel-Lucent, shares his view. “Without being too academic about it, in any interconnection model the largest players are the winners and perhaps have the least to gain from a global agreement, because they have the broadest base to negotiate bilateral agreements. So what exchanges do is level the playing field by lowering the cost of peering for the smaller operators for whom the cost of interconnecting is prohibitive.”

Whether the hub is an IP exchange/peering point, an Ethernet exchange or even a cloud-based platform, Modarres argues that this model enables the rest of the market to realise a larger footprint than they could otherwise, by being able to hand off traffic to points where they do not have network coverage. “So in terms of carrier confusion as to the need for IPX, I think it would be interesting to take the perspective based on their size, geography and footprint,” he says.

Whether the current contracts and mechanisms sustaining interconnection in the internet will be sufficient for a co-existence of traditional and future IP interconnection models is an issue that regulators and international industry bodies such as the ITU are still wrestling with. As such, it remains open for discussion and further exploration. “IPX is a great idea with a lot of support and many trials, but without the steadying hand of regulators, the larger players might dictate the shape of the future,” Ovum’s James concludes.