International Engineering Consortium
Web ProForums
Wireless Internet Network Communications Architecture

4. Orderly Migration: A Phased Implementation
The generations of cellular systems can be categorized by their data-handling abilities. Today’s second-generation (2G) systems are designed to support circuit-switched voice and rates up to 14.4 kps. The next-generation systems, including general packet radio service (GPRS) and IS–95B systems, can support packet service with data rates of 64 kbps or slightly higher. Trials for this generation are underway with such operators as France Telecom, Cellnet, and DDI in Japan. Third-generation (3G) wireless systems, which will be deployed in the year 2000, offer even higher-speed packet data services. Data rates for 3G systems will reach 384 kbps for pedestrian usage and up to 2 Mbps for fixed applications, opening opportunities for extensive wireless multimedia services (Figure 7).

Figure 1
Figure 7. Cellular Systems Evolution

As a quick recap, the reader now has a basic understanding of the functional components of the new communications architecture and a conceptual understanding of the cellular network evolution. The phased implementation discussion will now show how these various components can be gracefully added to existing networks for the following purposes:

  • extending the operator’s service offering to include high-speed packet data services
  • using the deployment of data services for establishing the basis for building up a core network
  • leveraging the core network for extending subscriber reach to other external networks and enhancing operator service-delivery opportunities

Phase-One Implementation
GSM and cdmaOne networks are circuit-oriented, and the service offerings delivered by these circuit networks have been focused primarily on voice and simple paging services. Beginning in 1998, there have been small deployments of data-enabling network elements—but in a circuit-mode rather than a packet mode. Packet mode reduces the dependency on the mobile switching center (MSC) for delivering data services; i.e., the data network connection is directly connected to the RAN, instead of being routed through the MSC. Enabling packet data services off the RAN and bypassing the MSC is the beginning step for separating the circuit-based world of the PSTN and the packet-based worlds of public data networks (PDNs) and the Internet.

Phase one for implementing the architecture begins with deploying GPRS or the public switched data network (PSDN) for enabling packet data services in GSM and cdmaOne networks. The alliance offers a complete GPRS solution that includes both the serving GPRS serving node (SGSN) and gateway GPRS serving node (GGSN) frame relay, asynchronous transfer mode (ATM), or some other link-layer transport capable of carrying IP packets. This wide-area, IP–based network is referred to as the GPRS data network and is regarded as a data overlay network. The SGSN plays the important role of interfacing with the RAN and mobility management elements of the GSM network—all of which have been upgraded in some respect for enabling packet data services in a circuit-based RAN network. A subscriber with a GPRS–enabled mobile station can initiate a packet data session via the RAN to the SGSN. The SGSN can route the data between it and the MS to another SGSN or to the GGSN for accessing an external data network. The GGSN is the IP gateway into external data networks, and it can deliver to the MS IP–based data services that are consistent with the look and feel of the Internet or other IP–based networks the subscriber may be accustomed to accessing via a wireline environment.

For cdmaOne networks, the alliance is offering the 3G PDSN for enabling packet data services. The PDSN can connect to the RAN locally to the base station controller (BSC), or—much like the GPRS implementation—the PDSN can connect over a wide-area, IP–based network using frame relay, ATM, or some other link-layer transport capable of carrying IP packets.

The phase-one implementation can be succinctly articulated as leveraging the deployed, IP–enabling, data services–network elements as the foundation for creating a core IP–based network. This core network will provide interoperability between RANs and connectivity to external networks and support the feature servers operators will deploy for increasing the number and variety of subscriber services. It will also provide the flexibility for fast deployment of new service offerings.

Phase-Two Implementation
Phase-two implementation can be planned by operators as a sequential step after completing phase one or as a parallel step to be implemented along with the deployment of phase one. In other words, as phase one requires the deployment of IP–enabling network elements, the operator can select data-enabling network elements that can also enable those functions associated with a core network. Data services itself will be a core network feature function representing a pooled resource for many RANs. This parallel phase-two implementation strategy helps the operator future-proof the network by planning the core network as a logical expansion of the data services network. The core network will be delivering IP–based services to the subscriber, and these services will include both voice and data applications.

Phase-Three Implementation
The synergy with phased investments already made begins to have its greatest impact for the operator in terms of reducing the network operations and provisioning costs and increasing new streams of revenue associated with the flexible creation and scalable deployment of new subscriber service offerings. Phase three is the build out of the core with internetworking gateways, packet voice call-control and signaling, and the new feature servers that will generate revenue streams for the operator.

The role of the circuit-based MSC will continue to regress in favor of the statistical gains delivered by packet voice and IP–based authentication, security, and mobility management. The circuit gateway will functionally migrate to the MSC for providing MS to PSTN or PSTN to MS call control. Digital signal processors (DSPs) on the circuit gateway will hold the vocoding algorithms for converting between a voice call encapsulated in an air interface frame and pulse code modulation (PCM). The voice-over–IP (VoIP) gateway function could be provided as an extended feature to the circuit gateway or the PDSN and the new 3G SGSN for universal mobile telecommunications system (UMTS) networks. The VoIP gateway will hold the vocoding algorithms converting between a voice call encapsulated in an air interface frame and an IP end point that may be an IP–enabled phone, enterprise IP–based private branch exchange (PBX), personal computer (PC), or any other voice-enabled IP device.

The feature servers play the role of service creation and service delivery, and the session manager working with the home service provider (HSP) provides subscriber authentication. Combined with the core network, the operator will be positioned to offer end-to-end, IP–based services and differentiate subscriber access to these services based on class of service, with class of service having quality-of-service (QoS) associations. A subscriber can be offered a menu of services and classes, and the operator can custom-tailor a service subscription to the needs of the individual while leveraging the savings and synergies produced by the whole of the subscriber population.

Broad Timing for Phased Implementation
All discussions of timing are ultimately qualified by the characteristics of the market that a specific operator is operating within and the business plans unique to each operator; e.g., new spectrum licenses are required for UMTS, wideband code division multiple access (W–CDMA), and personal communications system (PCS) bands. Therefore, timing will be discussed in a more general context of radio technology, geographic region, and what has been articulated by operators either directly or indirectly via industry research organizations (see Figure 8).

Figure 1
Figure 8. Projected Roll Out

In the context of radio technology, there are three 2G technologies that are creating a plan for offering 3G cellular services: GSM, CDMA, and time division multiple access (TDMA). Operators for all three technologies are considering delivering packet data services as follows:

  • GSM via GPRS
  • CDMA via the packet data serving node (PDSN) or IWU
  • TDMA via GPRS

GPRS packet data services will be deployed in network trials during the second half of 1999 with commercial service deployment anticipated by early 2000. GSM operators around the world are expressing strong interest in being the first in their markets to deliver GPRS packet data services. While many of the trials are in Europe (having the greatest density of GSM subscribers), the rest of the world’s GSM operators are equally enthusiastic.

CDMA packet data services share many of the same issues as GPRS with respect to the MS being the limiting element for delivering the higher-speed data rates promised by 3G radio. The general availability of CDMA packet data services will begin in Japan and Korea in mid-2000, while most North American operators will commercially deploy packet data services by the early part of 2001.

Earlier in 1999, TDMA operators elected to forego their own packet data services development in favor of adopting GPRS for its time-to-market advantages. The time frame for GPRS trials in TDMA networks is anticipated by early 2000. Commercial availability is expected during mid-2000.

Core network deployments in phase two are anticipated to begin with those visionary operators who already recognize the many advantages associated with creating a core network infrastructure. These operators clearly understand the core network as a prerequisite for implementing a 3G radio network and supporting the interoperability with 2G and other external networks. Core network request for information (RFIs) and request for proposal (RFPs) have already been issued, and vendor decisions will be made during the balance of 1999. For some operators, the core network is simply an extension of IP data networks already in place. For others, the core network is a new step away from the circuit-based networks they understand so well.

The build out of the core network with internetworking gateways and feature servers will begin immediately after the core is provisioned. The packet data gateways already discussed will become a pooled resource available to many RANs for delivering new classes of IP data services. VoIP and circuit gateway trials will be initiated, and, as call control for VoIP matures, so will the increase in voice traffic as an IP service with a correlated decrease in the voice calls traversing the MSC.

Registered Users
Enjoy exclusive access to free On-Line Education and receive the biweekly IEC newsletter.

IEC Newsletter
Get the latest industry information including critical insights from key industry leaders, technology briefings, and an Analyst Corner.
Current
Subscribe

Newsroom

IEC Corporate Member

Advertising Kit