Voice Over WiFi: Four Challenges, Countless Opportunities (Sidebar)
By Robert Sparks
The emergence of voice over WiFi (VoWiFi) broadband wireless stands as a logical component of a broader goal now shared by most buyers and sellers of communications: to liberate services from the limits of a single delivery system and make any given IP application or service available to paying customers wherever they are, via the best available delivery network at hand, and featuring emerging VoIP telephone calls.
WiFi is particularly attractive to IP applications users for its enterprise LAN-like speed and quality — virtues that make WiFi particularly accommodating to VoIP. As mobile manufacturers deliver handsets capable of accessing both WiFi and cellular networks, VoIP subscribers could soon begin to enjoy premium quality connections wherever they are available.
However, a handful of substantial hurdles must be overcome to make IP-based VoWiFi practical, available, secure, affordable, and easy to use. Most of these hurdles can be largely surmounted over the next 12 months, given the marketplace urgency to do so. As they are surmounted, VoWiFi will present opportunities to transform the ways people use networks and devices.
In the meantime, the industry must overcome challenges in four particular areas: Identity Sharing, Ubiquitous WiFi Network Access, Seamless Roaming Across Unlike Networks, and Device Capability Limits.
In VoWiFi, as with any other networks, users calling each other must be enabled to securely authenticate their identities to both service providers and each other. Yet WiFi’s highly diversified ownership and current dearth of inter-operator business relationships preclude any simple application of authentication methods or technologies from other types of access networks.
In traditional phone networks, subscribers delegate to the service provider authority to accurately identify any caller or ‘callee’ through the service provider’s own contracts to share that information with other service providers. This traditional caller implicitly trusts that “Caller ID” information is accurate, primarily because the service provider owns the physical drop into the subscriber’s premises and so is assumed to be the only entity using that connection.
However, because IP services are decoupled from physical networks, trust based on a physical network connection no longer applies. Encrypting presence, instant messaging (IM) and voice are the challenges. Which party is to take responsibility for making the assurance? What are the means to establish trust?
Web certification mechanisms provide one model through which third-party certificate authorities provide secure-site certificates based on trusted route. However, while this model works well with a limited number of fixed location servers, it is not designed to tackle a far greater number of individuals who are not tied to a single location. A VoIP phone will effectively need to attach a certificate to a person rather than a location.
Promising answers lie with the IETF’s current work on SIP Identity and SIP Certificates. SIP Identity allows a server or end point to make an assertion that is cryptographically strong — a solution. At the same time, augmenting the public key infrastructure (PKI) used by the Web, SIP Certificates would allow a subscriber to obtain a certificate from his service provider. In this scenario, the service provider knows that his subscriber is who he/she claims to be, and on that basis provides a certificate for the duration of an authenticated connection. Someone wishing to call this person securely can obtain this certificate from the service provider by subscribing to it using the SIP Events mechanism.
Ubiquitous WiFi Network Access
Currently, as end users move in and out of various WiFi network service areas, they are forced to manually log in to each new network. Such users engaged in a VoIP call would experience a dropped call, forcing them not only to dial again, but first to log in to the new network.
Cellular carriers have largely solved a similar cell-to-cell hand-off through a combination of home location register (HLR) technologies and provider-to-provider business arrangements, thereby enabling seamless national roaming across networks owned by different providers.
WiFi service providers will not be afforded a similarly elegant solution to provider-to-provider roaming, particularly because there are many more providers operating hotspots than there are cellular providers nationwide and worldwide. Pervasive and proactive business-to-business engagements will be more critical than technology in forging a common single-log-in solution.
Seamless Roaming Across Unlike Networks
Cellular carriers have proved that telephone call set-up signaling can be made to seamlessly hand off a call in progress from cell to cell for customers on the move. Such a seamless hand-off must be extended across unlike networks, if cellular subscribers are to be afforded the opportunity to enjoy higher quality WiFi connections wherever they are available.
Handsets must be able to detect WiFi services, and providers must make the signaling work. Building blocks do exist, but research on this remains incomplete.
Again the answer may lie as much with business arrangements as with technology. Because HLR technology does not apply to IP, carriers must work to define how a VoIP carrier may move a call across a cellular or WiFi boundary and what that will look like. Different segments of the call might be billed separately. Whatever mechanisms are applied behind the scenes, the key will be to make it appear to the subscriber that he/she is having one conversation.
Wireless industries are asking more and more of portable handset devices that are compute and power limited. Providing identity certificates and real-time detection and analysis of networks will challenge the current generation of devices. Despite promise at the high-end of these devices, the average person on the street will demand an affordable device with more power than is available to date.
Handset manufacturers will have to meet this challenge by wrestling with the tradeoffs between central processor unit (CPU) cycles and battery time. The technology must be evolved to provide the kind of portable computing these applications will demand by delivering the greatest amount of CPU cycles at the lowest consumption of the battery and at the least cost to users.
On The Other Side
As the industry breaks through these barriers, it also must develop and standardize protocols to leverage a particularly promising opportunity: integration of presence to empower subscribers to know a person’s availability, capabilities, and environment.
Action choices multiply through the knowledge that someone is on the network but in a meeting, or on a handheld device rather than a laptop with a multimedia coder/decoder.
Such capability status might be indicated by a power bar similar to signal strength indicators common to today’s cell phones. Through such knowledge, users can be afforded choices to accept or defer a call, to talk with a client over an optimal or sub-optimal connection, to click or not click on an ad for information on a local restaurant. A presence server might auto-conference board members, family members, or members of a fan club who have been waiting for a moment of common availability to discuss a decision.
With presence-enabled VoWiFi, real-time interactive communications will expand in dimension, challenging service providers to develop and package presence capabilities that appeal to — and raise the expectations of — business users, consumers, advertisers, and other stakeholders in their services. Raising and meeting these expectations lies within the industry’s grasp.
Robert Sparks is the CTO for SIP softphone provider Xten Networks, Inc. For more information, please visit the company online at www.xten.com.
[ Return To The January 2005 Table Of Contents ]