New Architectures for the HSDPA Handset Revolution
Nigel Toon, vice president and founder, Icera Inc, looks to the future of the mobile industry and explains how the handset baseband architecture will radically change as it is driven by new high-speed cellular broadband services enabled by HSDPA.
The days of cramming handsets with countless features are numbered, with traditional platform architectures already showing signs of strain. The call for a new approach is being answered by the launch of HSDPA in 3G networks and the advent of cellular broadband, which looks set to revolutionize the mobile industry.
Consumers will now demand more specialized and user-friendly devices that deliver a complete service for specific functions. Over the next three years, the shift to HSDPA and HSUPA will reshape the mobile handset landscape to address this demand, driving new classes of mobile devices and services, with significant implications for handset and baseband design.
Voice Still #1 Despite Handset Evolution
When GSM took off in the mid-90s, handset architectures were focused on delivering voice, a single function with a superior end-to-end service. Successful early GSM devices delivered an intuitive user interface and extended battery life in a small form factor, with color displays becoming widespread some years later and the first camera phone appearing in 2000.
Today, most mid-range phones adopt the ubiquitous ‘one size fits all’ approach, sporting large color displays, cameras and, increasingly, music playback facilities. Although these new features have been important in driving handset upgrades, voice still dominates in terms of usage and revenues. Fashion aside, voice is still the main feature for which consumers buy their mobile phones, as evidenced by the remarkable success of handsets with cutting edge designs but average feature sets.
Future Trends – Connected Consumer Devices
A new environment for mobile devices will be created by the introduction of ubiquitous, high-speed cellular broadband data services enabled by HSDPA. Connected consumer devices will be the new trend in wireless handsets, although the traditional voice terminal will continue to evolve.
Consumer devices have undergone dramatic change in the last five years, both shrinking in size and growing in function and quality. Solid state storage has largely replaced tape and film. Laptops, camcorders, music players, gaming devices, digital cameras and mobile video players can be found in most households. Devices are beginning to offer WiFi connections for file sharing and synchronization.
The laptop is the first product to benefit from cellular broadband. The arrival of HSDPA datacards will be quickly followed by PCIe Mini-Card modules embedded as a customer-build option inside the laptop to enable a seamless broadband connection wherever it is used.
Next we will see media players using HSDPA modems to deliver always-connected service, allowing music or video files to be selected either from on-board solid state storage or as new downloads from a portal service. The user interface will make this process seamless. No need to connect and then search through layers of menus – a simple menu click will connect users to new content; once selected, a key press will initiate download and billing. Full track audio download will be possible in approximately 30 seconds, compared to the three or four minutes taken with R99 WCDMA.
HSxPA will enable interactive play and new game downloads for portable gaming devices. The same type of end-to-end service will also benefit digital cameras and other consumer devices. Mobile phones will follow the same trend as these devices by specializing on a single function and delivering a competitive consumer experience. Indeed, major manufacturers are already introducing media- and business-focused handsets. One of the major items for business users is push email – complete with attachments, this will become a reality for users of cellular broadband through HSDPA.
Figure 1. Conventional architecture – integrated LSI baseband and applications processor
Although it seems economical to integrate the baseband and applications functions in every case, the proliferation of connected consumer devices that serve different specific markets means that it is no longer possible to cover a full portfolio with just one economic platform. The challenge facing the handset manufacturer is how to efficiently address the myriad different handset capabilities, operator-specific requirements and evolving standards while addressing a global market.
Push email is a good specific example. There are several push email systems from different companies (e.g. RIM, Good Technologies, Intellisync), each of which requires a server within the operator or corporate network, and none of which is compatible with another push email system. This means that the handset manufacturer will need to integrate different email clients for different target customers or operators, and that requires additional work and customization.
There are also several music services to which operators could be specifically aligned: Real, Apple iTunes, and Microsoft Windows Media to name but a few, and they use a variety of digital rights management (DRM) technologies: OMA, Apple Fairplay and Windows Media, for example. The choice of service will vary from operator to operator. It will also vary according to customer preference: consumers may prefer to use the same service on their handset as the service they use at home.
Ideally, the system architecture will provide sufficient flexibility to support a number of application alternatives for services such as push email and DRM. An alternative architecture, found more frequently in high-end devices, uses a separate applications processor to deliver more advanced applications functionality and to allow the baseband and applications to evolve at different rates in differentiated devices.
Figure 2. Separating baseband and applications in a high-end multimedia architecture
Now a debate is raging over which architecture is best for wireless devices in the future cellular broadband space. Chip packaging technology makes it possible to achieve the same form factor from both architectures, and costs are comparable, so the difference comes down to cost of ownership of future devices through flexibility and design reuse. If, to maintain technological and performance leadership, it is necessary to design a new architecture for every new product, then there are no economies of scale through a platform; each product is bespoke, leading to high engineering costs, high stock risks and longer time to market.
Baseband Challenges: One Core or Two?
Addressing all of the above is a major challenge for the baseband silicon. Most R99 WCDMA basebands use a single RISC processor for the protocol stack, and baseband manufacturers are now trying to evolve their R99 WCDMA solutions to address HSDPA.
Enabling 3.6Mbps and 7.2Mbps HSDPA is a significantly more demanding task for the processor than enabling 384kbps WCDMA. In reality, HSDPA will consume about 50 -150 MIPS (peak demand) of the typical 200 MIPS RISC processor for the protocol stack alone. To achieve this performance will require a significant amount of memory (between 100KB and 256KB), tightly coupled to the processor core (TCM) for data management. The external memory bus can be up to 50 percent loaded.
With all of this processing and memory required just to deliver HSDPA, it is clear that a single RISC core will be insufficient to address both operator and customer demands for HSDPA handsets. The spare processing capability would not be sufficient to support multitasking applications in parallel with HSDPA download and would clearly not support an open or semi-open operating system. This means that a multicore approach is required to address the requirements for connected consumer devices.
The applications function that is tracking the changing applications environment will be separate from the baseband processing, and focused on delivering the highest level of performance for a specific function at the lowest power and cost levels. The communications function will be delivered as a small, integrated device that can be easily interfaced and can deliver the multimode cellular and WiFi connectivity functionality. The compute plus modem architecture will dominate the handset, however, with different scaling and specialization on the compute function being combined with a flexible multimode wireless modem.
Wireless Soft Modem Solutions
Icera is focused on enabling cellular broadband-connected consumer devices. The company has recently introduced the world’s first wireless soft modem that delivers a complete HSDPA multimode modem that is upgradeable through software to support HSUPA and other wireless standards such as WiFi and mobile digital video in a single low-powered device. This approach delivers enhanced time to market, low power and enhanced network capacity – key benefits for future handset development.
A number of companies are focused on the consumer-centric compute SoC devices, and partnerships are evolving that will deliver complete platform solutions for these new connected consumer devices to provide whole-product support for the next generation of mobile handsets.
©2010 Synopsys, Inc. Synopsys and the Synopsys logo are registered trademarks of Synopsys, Inc. All other company and product names mentioned herein may be trademarks or registered trademarks of their respective owners and should be treated as such.