IP transformation is the migration by service providers to a packets-witched next-generation network (NGN). The generic NGN model is based on a horizontal platform that means separation of transport, control and service layers. An IP/MPLS backbone is used for transport, while the control and service layers typically comprise a loose collection of voice over IP (VoIP) network elements, such as softswitches, voice application servers, media gateways and SIP-capable endpoints.
Softswitches are employed to deliver software-based call control, while media gateways enable transport interworking between TDM and packet-switched networks. Voice application servers reside in the application layer, supporting IP voice and multimedia services, and communicate with softswitches using a session initiation protocol (SIP) interface. SIP is the signalling protocol used for creating session-oriented connections between two or more end-points in the IP network, and is also a key enabler in the GSMA’s IPX. SIP allows the exchange of traffic such as VoIP, IP video telephony, presence and messaging.
Although the core SIP protocol is defined within the Internet Engineering Task Force (IETF), its development has been fragmented by numerous subsequent drafts, and in-country and vendor-specific implementations. Other VoIP protocols currently in use include Bearer Independent Call Control (BICC) and H.323. In order to address this fragmentation and support interworking with the PSTN, IPX has adopted the use of SIP-I, which is SIP with encapsulated ISUP protocol. ISUP is part of the SS7/C7 signalling suite used in the PSTN.
A gateway to IP
The need to support existing TDM infrastructure, and interwork services such as VoIP and traditional TDM voice, calls for a “next-generation” approach to interconnect. Investment is focussed on session border controllers (SBCs), which control and manage real-time multimedia traffic flows between IP networks and, as their name implies, provide the border between IP networks for carrier’s carriers (peering borders), as well as the border between retail service providers and their customer’s premises equipment (access borders).
SBC equipment vendors report a pick-up in activity in the last six months. Acme Packet has now shipped some 10,000 SBCs to more than 1,000 customers in 100 countries. “We are supporting transitions to IMS, Rich Communications Suite (RCS), and voice over LTE – all of which moves towards SIP-based next-generation communications,” says Kevin Mitchell, director, solutions marketing, at Acme Packet.
According to Natasha Tamaskar, VP of product marketing at Genband, the initial target market for interconnect SBCs was wholesale carriers. She says: “Typically, this would be the smaller service providers going with simple wholesale models, whereas the larger operators are looking at IMS as a means of enabling interconnectivity. However, IMS is still a market driven by less than 100 service providers.” Tamaskar says that a growing number of carriers are now targeting multimedia connectivity. “Multimedia brings in all types of requirements around transcoding, since you have all the different types of IP protocols connecting on the signalling side, together with the fact that there are multiple types of media connecting as well.”
Migratory alternatives
There are two routes to migrating traditional TDM interconnect to IP, and carriers will have different requirements depending on their size, network configuration and choice of technology. For example, an IMS-centric carrier will typically go with a decoupled architecture, where there is a control plane, and a media gateway, with the latter handling the TDM-to-IP and IP-to-IP transcoding for interconnection. This is required by carriers interconnecting heterogeneous networks. However, many carriers do not need this and therefore opt for a coupled architecture via the SBC, which is an all-IP platform and has an integrated control and media plane.
Both approaches have their advantages. With an integrated solution, there is just one point of interconnect, so all traffic can be controlled effectively. In the decoupled architecture, the control plane can be deployed centrally, while the media plane can be deployed at the network edge, so less traffic has to be routed via the core network.
“It is really a matter of how the carrier has configured its network. There is no good or bad way,” says Tamaskar. “And with transcoding, some carriers will prefer a centralised approach because they can figure out how to grow in a better way and manage the associated transcoding requirement. Others will prefer to handle transcoding at the edge of the network. Smaller carriers in particular prefer to keep things simple with an integrated solution that is centralised.”
Regardless of the approach, efficient routing is essential in handling the complex requirements of carrier interconnect. “Otherwise within a matter of minutes, the carrier providing the interconnection could be losing a lot of money,” says Tamaskar, who believes the adoption of IPX will take several years. “Service providers worldwide are looking for a smooth transition from very mixed types of connectivity, to a model of all-IP, and this is where the standards come out of the i3 Forum.”
Converging the standards
“The challenge we have for voice is that IPX is based on the GRX mobile data model, rather than being from the perspective of transitioning voice to an IP environment,” says Christian Michaud, senior VP at Tata Communications and a member of the board of both the IP Interworking Alliance (IPIA) and the i3 Forum. “The IPIA has been working to adapt that model in terms of the technology, the business model and the way that we manage the voice business.”
Michaud says much of this work was conducted in collaboration with the i3 Forum and the IPIA, before engaging directly with the GSMA. “The model has evolved to the point now where we have an IPX that is applicable to all aspects of the migration to IP – whether in terms of carrying content, data, signalling, push email connectivity or voice.” New applications are also now being worked into the IPX framework, such as high-definition voice and content downloads.
Standards development for IP-enabled points of interconnect includes: IPX; Speermint – Session Peering for Multimedia Interconnect, an IETF working group defining the architecture to identify, signal and route delay-sensitive communications sessions for Layer 5 SIP peering; and IMS – which provides a blueprint for telecoms operators to build and launch IP applications, converge fixed and wireless networks, and support seamless user roaming, though standards development is ongoing and has yet to address open interoperability in detail.
A further key standard to enabling interconnect is the IETF’s electronic numbering (ENUM). ENUM is a specific service within the domain name system (DNS) that automatically translates a telephone number into an IP-based address known as a universal resource identifier (URI). This simplifies interconnectivity by affording service discovery to the originating network. There are three types of ENUM: end-user or public ENUM which relies on end-user opt-in and provides an open database; private ENUM – implemented within an operator for internal resolution and service discovery; and carrier ENUM which provides the benefits of improved routing and interoperability, with private inter-working ENUM services between operators with peering agreements.
Capacity and management
The explosive growth in IP traffic provides a further challenge for carriers as they migrate to NGNs. Cisco reports that annual global IP traffic is expected to quadruple over the next four years and to exceed three-quarters of a zettabyte (767 exabytes) by 2014. “All service provider and carrier networks face the explosion in bandwidth demand, yet the trend towards flat-rate pricing means they will see very little return in additional revenue,” says Joerg Ammon, systems engineer at network solutions provider Brocade. “The key for carrier’s carriers is to use a technology that can grow with the traffic and is simple to operate. A single 100Gbps interface is much easier to maintain than 10 such interfaces. Transparency to monitor the traffic and report on QoS and SLAs delivered is also essential.”
The use of MPLS at the transport layer is another weapon in the armoury of carrier’s carriers. “Management of capacity and planning is an ongoing challenge,” says John Wick, VP network services at Syniverse. “MPLS enables network operators to evaluate different streams of traffic across the common network and prioritise packets for efficient delivery. This allows the creation of end-to-end circuits across any type of transport medium while providing the management tools to test, troubleshoot and maintain a highly reliable, flexible network.”
Interconnect models such as IPX pose a significant challenge in terms of billing, rating and mediation. “The sheer volume of data being generated means there will be a trade-off in tracking that data and whether the expense of capturing it is justified,” warns Alastair Hanlon, who has global responsibility for business development and marketing at Convergys. “One of the key challenges posed by IPX for carrier’s carriers is support of the cascading payments model, which is designed to ensure that everyone in the value chain receives a share of the revenue. This requires co-ordination of all call detail records and settlement, which in an IP world calls for massive scale and flexibility in supporting multiple charging models for multiple service types.”
Convergys is researching whether it is possible to aggregate usage and keep account of the transactions as opposed to the detail of each transaction, which would significantly reduce the volume of data generated. Says Hanlon: “There will always be some arrangement, but you might get to the point where the revenues carriers are getting from carrying the traffic is so little that they go to a model where they charge a flat fee so they do not need to count it any more.”
