You know how the old sayings go:

"The quickest way to make a car run fast is to take the brakes off it and run it on methanol."

"The easiest way to make grandad run fast is to take his pants off and set his arse on fire."

Now startup Starent Networks Corp. is looking to add a "new" phrase to the old chestnut lexicon: "The easiest way to add fancy-shmancy services like push-to-talk (PTT) to older CDMA and GSM wireless networks is by using Starent's service node technology."

OK, so it's not very catchy. But that doesn't mean Starent's products won't catch on [ed. note: geddit!?] with service providers. Push-to-talk is a technology that enables subscribers to use their phones like walkie-talkies, quickly connecting with other users for short conversations with more than one person at once. Subscribers like it, which means operators of skanky "it's seen better days, I can let you have it for a fiver, guv" network infrastructure need to find a way to roll it out.

Nextel Communications Inc. (Nasdaq: NXTL) is currently the only operator in the world that has a commercial push-to-talk offering, courtesy of its proprietary iDEN network (see Nextel's Nationwide Walkie-Talkie). Other operators in the U.S. and Europe are racing to get PTT services up and running on their 2.5G general packet radio service (GPRS) and CDMA2000 1xRTT networks (see Walking, Not Running, to PTT, Winphoria Talks Back and Vendors Push Carrier Upgrades).

Starent, which has just signed an equipment deal with Brazilian carrier Vivo (see Starent Wins Vivo Deal), points out that rolling out PTT only on newer networks limits the potential revenue that operators will see from such a service, simply because many users will not upgrade to a new phone, and so never get to use PTT.

Using its service-node technology, Starent has come up with a way to implement a pseudo-PTT service that will work on older phones and networks.

One of the main problems with implementing PTT on GSM-based and CDMA networks is that it that it can take time to set up "data calls" over cellular networks. There can be seconds of delay before a user is connected to the subscriber they wish to talk to (see Ericsson: Pain-to-Talk?).

Starent has not come up with a miracle solution to call-setup problems on these networks [ed. note: so we don't get to shout "Hallelujah, Brother!" then? What a pity]. In fact, the company has simply sidestepped the problem. The operators can set up numbers for users to call when they want to initiate a PTT session with one or many users (which is where the technology really differs from a standard voice call). Once a user has started a session, they remain directly connected to another user or group for as long as they continue the call.

Starent handles the VOIP side of the equation by routing the call through its Packet Data Serving Nodes (PDSNs) or GPRS Gateway Support Nodes (GGSNs) -- service nodes (wireless routers, basically) that connect the radio elements of a wireless network to the IP core and handle tasks like traffic management and QOS levels (see What's the Wireless Router Market Worth?). "We essentially treat it as a data call," says Gennady Sirota, VP of product management and marketing at Starent.

Gennady says carriers with a large installed base of subscribers who aren't likely to be tempted to upgrade their handsets are the primary target for such services. "Like Vivo, like China Unicom," he says.

Starent isn't alone in talking up the service-creation potential of its box. Startup rivals, especially Megisto Systems, have become increasingly focused on promoting the extra services that their boxes can enable, even if the equipment ends up alongside an existing vendor's service node.

— Dan Jones, Senior Editor, Unstrung