The debate over dual-mode data

At Supercomm, BellSouth announced it would kick off cellular/Wi-Fi convergence trials in Atlanta. Shortly afterward, BT launched its much-anticipated Bluephone service, which uses a Bluetooth connection to offload traffic from the cellular network. Back across the Atlantic, SBC Communications has made much ado about its planned launch of voice-over-Wi-Fi services in 2007.

That's a lot of activity around convergence, but all of that activity has one thing in common: voice. For all the talk of integrating the local area and personal wireless networks into the wide area cellular footprint, discussions have revolved almost entirely around voice — cutting down per-minute voice costs, offloading voice minutes onto the IP network and using Wi-Fi as a picocell voice technology. Despite the fact that voice over IP (VoIP) and unlicensed mobile access (UMA) solutions are merely two IP applications among many, the use of Wi-Fi or other local wireless access technologies to offload IP traffic from the cellular network simply hasn't garnered much attention.

The main reason is the lack of traffic on cellular data networks today — carriers don't have anything to offload yet. And even when carriers do manage to load their data networks, using Wi-Fi to cannibalize traffic from their expensive 3G networks isn't their first inclination. But not every carrier has the investment in an UMTS or EV-DO network that the Tier 1 carriers do — several are looking for ways to leverage the vast broadband data infrastructure wired in the ground. And over in Asia, the first signs of network crowding are starting to reveal themselves.

“In Japan and Korea, you're starting to see the saturation of data networks,” said Mike Mulica, president and CEO of BridgePort Networks. “Real-time downloads of video have become so popular they are taxing carriers' 1X RTT EV-DO networks. To handle the demand, carriers there are requiring customers to download programming in the middle of the night to watch the next day.”

In the U.S., where mobile data usage is still in its infancy, you'll be hard-pressed to find any facilities-based wireless operator looking to offload cellular IP traffic onto a Wi-Fi hot spot or Bluetooth connection, Mulica said, but as data traffic increases, you'll see some carriers branching off. Providers with broadband last-mile solutions, like Sprint and T-Mobile, will invest heavily in their wireless broadband services, seeing it as their primary solution for data access, both on the wholesale and retail fronts. But providers owned or partially owned by an ILEC, like Cingular and Verizon Wireless, will be tempted to use their extensive wireline resources to build very robust and ubiquitous data networks.

The most freewheeling of providers, however, will be the mobile virtual network operators (MVNOs), Mulica said. With little or no investment in network infrastructure, it makes no difference whether their users are accessing their services via a fancy 3G network or over a Wi-Fi connection at Starbucks. Mulica pointed to entrepreneur Sky Dayton whose SK-EarthLink venture will be one of the first data-centric MVNOs in the U.S. Faced with the relatively low cost and high capacity of Wi-Fi and the high cost and low capacity of the wide area cellular network, Mulica said, SK-EarthLink will look to use as much Wi-Fi as possible, which Dayton recently confirmed.

“Once you get out of a hot spot, 3G is a better network, but when you're home or in a densely populated area, Wi-Fi is often the better option,” Dayton said in an interview with Telephony sister publication Wireless Review. “Neither one is a complete solution on its own. So the question becomes, how do you put them together?”

While there has been intense focus on UMA and other cellular/wireless LAN (WLAN) voice convergence solutions, dual-mode data hasn't generated the same interest. In part, the reason is market demand. While wireless data has made inroads worldwide, it's still not the critical application that cellular voice has become. Customers are looking to drop the per-minute cost of cellular calls, especially in areas where they already have infrastructure in place, like the corporate enterprise. Carriers, while not so keen on seeing their per-minute charges fall, are still keen on finding ways of getting into high traffic zones like the aforementioned corporate headquarters without deploying new cellular infrastructure.

Wireless data, on the other hand, is still not on the radar screens of the majority of customers, especially the consumer. While there are enterprises that rely heavily on wireless data, they primarily fall into the vertical market category, where proprietary, tailor-made solutions specific to the customer prevail. For the general mobile professional, wireless data is handy, though not yet a critical application. But as carriers strive to make mobile data a critical application, the last thing they want to do is undercut their margins by dumping valuable data minutes onto a Wi-Fi or Bluetooth network.

UMA and wireless VoIP handsets may be getting all of the attention today, but a handful of smaller vendors are looking at data convergence, anticipating a greater demand for data handoff in the future. Although BridgePort is now focusing on mobile VoIP solutions, it has built its network convergence gateway to be indifferent to IP packets, whether they carry voice packets or plain-old bits of data.

While UMA solutions are finding ways to take the GSM transmission and tunnel it through the IP network, BridgePort's gateway is entirely session initiation protocol-based, tracking an IP address attached to a mobile device regardless of whether it's traversing the Wi-Fi hot spot, corporate LAN or wide area cellular network. According to Mulica, the gateway basically registers the Internet as a large roaming cell. Just as a GSM data customers in the U.S. would roam onto another GSM carrier's GPRS or UMTS network in Europe, a customer could roam from a GSM network to any Wi-Fi network. From the carrier's back-office perspective, the customer is merely in a roaming partner's footprint.

Meanwhile Azaire Networks is tackling the data convergence issue head-on. Its IP Converged Network platform is purposely built to meld GPRS/UMTS networks and the WLAN, and the vendor is already starting to attract some interest in its solution in Europe. At CeBIT in March, Azaire and T-Mobile demonstrated a continuously running virtual private network connection passed off between a T-Mobile UMTS base station and a Wi-Fi hot spot. And Mobilkom Austria has selected Azaire's solution to launch a data convergence solution utilizing both its UMTS and hot spot footprints.

Azaire's Metro Services Gateway essentially threads IP traffic from a Wi-Fi access point directly to a carrier's gateway GPRS support node, making a hot spot look exactly like another node on a carrier's network. Its services control node, meanwhile, acts as the authentication and management interface regulating the handoff between networks and managing the complex billing back and forth between wireless providers and hot spot roaming partners. The architecture creates more than just a convenient dump for broadband packets off of the cellular network, it's able to replicate a carrier's service almost exactly. It easily handles standard Internet payloads like streaming video and e-mail, it can route short message service and eventually advanced services like multimedia message service or push-to-talk over Wi-Fi access.

“Carriers are looking to pull Wi-Fi into their networks — and not just for voice,” said Bill Howe, Azaire president and CEO. “All of these carriers have broadband pipes, but they are discovering they are very inefficient for delivering broadband applications over a macro-cellular network.”

The one basic assumption driving any IP data convergence strategy is that the per-bit cost of pumping data through a Wi-Fi or Bluetooth network is cheaper than piping it through a 3G network. At least one vendor takes issue with that assumption. Novatel Wireless CEO Peter V. Leparulo said that constant capacity improvements to core 3G technology and the rapidly falling price of radio access infrastructure may cancel out Wi-Fi's budgetary appeal.

Incremental improvements in 3G, such as high speed downlink and uplink packet access (HSDPA and HSUPA) as well as future releases of CDMA2000 EV-DO technology are expected to push network bandwidths up from a few hundred kilobits per second to the multi-megabit range. Furthermore, as more 3G infrastructure is deployed, the cost of the equipment is shrinking, especially as vendors target carriers in developing countries where ARPUs are significantly lower but addressable populations are extraordinarily higher. The end result, Leparulo said, is that a carrier with an overloaded network might find it just as cost-efficient to deploy another cell site than to offload packets onto Wi-Fi, giving it the added benefit of keeping the customer on the provider's network.

“Base stations are becoming much more commoditized,” Leparulo said. “When operators find they need more capacity, it could be just as easy to drop in another base station than migrate their traffic to another network.”

Leparulo expects that 3G networks will become so robust that UMTS in some cases will be used as backhaul for Wi-Fi networks. In fact, he envisions a UMTS backhaul scenario as a perfect way for carriers to deal with issues of congestion where it's not feasible to add another base station. By deploying their own Wi-Fi access points, carriers can use Wi-Fi to connect multiple users to a single UMTS transmission. The architecture would create a single UMTS data session back to the base station, a much more capacity-efficient method than managing multiple UMTS sessions. Carriers could take advantage of some of Wi-Fi's benefits without ever leaving the physical 3G network, Leparulo said.

However, the issue of whose network the customer resides on may soon be a moot point entirely. With IP multimedia subsystem (IMS)-based networks coming online, focus is expected to shift from the access technology to the services themselves. Right now, customers don't care if they're on a Wi-Fi or 3G network as long as their applications work properly. If IMS has its way, carriers won't care if customers are on their 3G networks or Boingo's hot spot footprint so long as the services work properly.

IMS, however, is still a long way off, both as a fully realized technology and as a concept many in the industry are capable of wrapping their minds around. Data as well as voice convergence between cellular and Wi-Fi is a way of immediately bridging the gap, its proponents claim.

“IMS is stalled,” BridgePort's Mulica said. “Carriers are not willing to make the investment until they see the returns. This BridgePort approach will be adopted as a way to justify IMS.”

Regardless of whether cellular/Wi-Fi convergence is the missing link between the current and next-generation architecture, carriers are still on the fence about the technology. They must protect their investments in their new high-speed data infrastructure but also meet customer demands for cheaper services, which means cheaper access alternatives. It's not an easy rift to mend, and according to Novatel's Leparulo, carriers are expressing two distinct and contradictory views about the technology.

“Some of them are afraid of it because it will take traffic off their networks,” he said. “Others are for it — because it will take traffic off their networks.”