The anticipated deployment of High Speed Downlink Packet Access (HSDPA) technology by wireless service providers ushers in a new era of innovative and competitive mobile applications. This is a dynamic era, where the number and variety of compelling applications will fulfill the potential of wireless networks that was envisioned by service providers when they first crossed the line from analog to digital technology in the mid 1990s. But just as HSDPA empowers more dynamic applications, it also demands more dynamic testing on the mobile device to ensure a smooth service launch and provide the high levels of reliability and quality consumers have come to expect.
In migrating to second- and third-generation wireless standards, test engineers largely relied on standards-based conformance testing. In implementing a 3G WCDMA RF test plan, for example, engineers used test cases and procedures that conformed to the Release 99 specification to perform static testing.
Our experience in testing wireless devices during the migration from GSM to WCDMA networks provides several useful insights. Two WCDMA handsets might both have passed test parameters defined by the Third Generation Partnership Project (3GPP), but they often varied greatly in performance and in the demands they made on network resources. This is because the 3GPP defined minimum performance tests to ensure that 3G would work, but it left it up to the service providers to ensure that their networks operated efficiently and profitably with specific end-user devices.
We also observed that the introduction of competing CDMA EV-DO (Evolution-Data Optimized) services faced similar obstacles. Here too, engineers found that test standards weren't always comprehensive. The early revisions of new standards were based on visible needs and were written without the benefit of real-world experience.
Both technology rollouts revealed a “Conformance Gap” in testing procedures. They underscore the need for engineers to go beyond just conforming to the spec and be proactive in testing the strengths and limitations of the wireless device. Relying only on static testing scripts may leave the service launch vulnerable to failure in the real world.
Furthermore, the complexity of the HSDPA standard is far greater than the basic Release 99 WCDMA protocol. An overdependence on letter-of-the-law conformance testing widens the gap even further and leaves the carrier exposed to the negative consequences of bad user experiences with the latest 3G wireless devices.
HSDPA is a central part of the Release 5 UMTS specification and is sometimes referred to as 3.5G technology. It promises packet-data service with data transmission speeds up to 14 Mbits/s to support multimedia services. The resulting competitive advantages that the protocol offers service providers are so great that HSDPA deployment schedules have been accelerated across the worldwide UMTS community. But for HSDPA to succeed in the market, user devices must be thoroughly tested for efficient and proper operation.
To meet this challenge, engineers across the value chain must quickly rethink their testing procedures to ensure a successful HSDPA service launch and maximize the benefits of this new technology. Dynamic testing can fill the conformance gap by employing extra testing steps that more closely replicate real-world 3.5G wireless networks. Dynamic testing raises the bar from static conformance to true performance.
Higher levels of testing
There are significant differences between Release 99 and HSDPA specifications. HSDPA is far more complex and demanding. These differences further illustrate the need to go beyond conformance testing. Engineers are still grappling with the complexities of the Release 99 spec, the code domain environment, and the use of power control to provide "soft" network capacity control. In Release 99, power control is a key factor in the success of a network deployment, as it affects both the cost per user and the network's subscriber capacity.
HSDPA adds to the complexity by employing a different means of link adaptation control. It manages the data rate available to the user and relies on low-level (physical layer) feedback from the device to ensure success.
Release 99 devices primarily use the Dedicated Channel (DCH), where one user is provided a single downlink code channel. In contrast, HSDPA relies on the High-Speed Downlink Shared Channel, or HS-DSCH, to provide user data. This channel is shared between multiple users and is controlled by rate adaptation based on time-scheduling at the Node-B base station. But because there's no standard to define how the Node-B implements this scheduling, an effective test plan requires a certain amount of flexibility.
WCDMA uses orthogonal variable spread factor (OVSF) codes to separate users in the code domain at the physical layer. Different from Release 99 DCH channels, the HS-DSCH uses multiple parallel OVSF codes, each with a fixed spreading factor of 16. HSDPA lets one user simultaneously employ several parallel codes to transmit data at very high speeds.
HSDPA service also relies on the Hybrid Automatic Repeat reQuest (HARQ) in the Medium Access Control-High Speed (MAC-HS) layer, which in turn depends on the Uplink Dedicated Physical Control Channel for HS-DSCH (HS-DPCCH) and up to one of four High Speed Shared Control Channels (HS-SCCH) for control (Fig. 1). The HS-DPCCH is also responsible for sending Channel Quality Information (CQI) from the user device to the network, providing yet another parameter to allocate resources (Fig. 2).
1. HARQ and CQI statistics are a critical metric to assess performance and efficiency of wireless data traffic.
2. The HSDPA network analyzes the CQIs to dynamically allocate bandwidth.
Coding and modulation can vary. Adaptive Modulation and Coding (AMC) introduces a host of possible test permutations, as well as new implications imposed by higher-order (16QAM) modulation. HSDPA also must work seamlessly with existing Release 99 and GSM service. This means proper operation of Release 99 voice services alongside HSDPA data services. A comprehensive test plan must address all these factors, and include tests defined by the 3GPP as well as tests beyond the scope of the 3GPP standards.
The 3GPP standards community is working toward defining RF and protocol test cases for HSDPA, defined in TS 34.121 and TS 34.123, respectively. For Release 99, these two documents form the basis of User Equipment (UE) certification criteria as defined by the Global Certification Forum and the PCS Type Certification Review Board.
HSDPA is still a new technology and the volume of test cases is much small compared to Release 99. The standards community is trying to add tests. Once these test cases are created, a basis for RF and protocol performance testing will be established. However, several key aspects of HSDPA are outside the scope of traditional conformance testing. To build a truly comprehensive HSDPA test plan, the real-world environment and network interoperability scenarios must be factored in as well.
In TS 34.121, six RF parametric conformance test cases are specified in Sections 5 and 6 (as of v6.3.0). These tests evaluate basic UE RF transmitter and receiver functionality. The TS 34.121's Section 9 is titled "Performance Requirements for HSDPA," but more accurately addresses HSDPA minimum performance testing. This section provides a starting point for a more comprehensive HSDPA performance test plan.
The specification only identifies one fixed transport block size for each UE category and modulation type. Also, it's assumed that there are no other active HSDPA users. In the real world, the UE will encounter many different transport block sizes. A given UE's service provider is highly dependent on the number of users competing for shared network resources. Hence, the rate at which a UE will be allocated HSDPA resources continually varies.
In addition to RF tests, the 3GPP RAN5 also defines a protocol conformance test specification in TS 34.123. While TS 34.121 and TS 34.123 offer a baseline for HSDPA testing, there are several key areas that aren't addressed. HSDPA's salient feature allows for fast allocation of a shared resource. Therefore, real-world testing requires an environment that's dynamic both in terms of RF conditions and in having other users compete for resources.
HSDPA will be overlaid on top of parts of Release 99 networks so handovers to and from Release 99 will be common. In general, mobility for HSDPA is defined differentlysoft handovers aren't supported. Because the MAC-HS function is located in the Node-B, a handover from one Node-B to another requires a MAC-HS reset and reliance on higher layers to maintain the connection.
To ensure that the HSDPA technology delivers on its promises, network operators and UE manufacturers must be thorough in planning detailed test methodologies. These plans may start with testing defined by the 3GPP, but must extend far beyond these minimum requirements. Failure to do so will produce inconsistent high-speed service, resulting in an unhappy subscriber base and great cost to the operators and device manufacturers.
About the author
Salim Manji is a product manager at Spirent Communications in Eatontown, N.J. He earned a PhD from Rutgers University's Wireless Information Network Laboratory. He can be reached at firstname.lastname@example.org.