Tuesday 16 December 2008

Metro Femtos another option for LTE.

Around six months back, Unstrung article mentioned that Vodafone dreams of Metro Femto. Now Doug Pulley of picoChip s all set to champion this concept. In the recently concluded LTE World Summit, he said "The macrocell is dead. It's a fallacy to think you can reuse existing cell sites to get LTE services. That whole premise is broken."

Both China Mobile and T-Mobile have said they plan to use existing 3G cell sites for their LTE networks. But Pulley contends a traditional macrocell deployment won't work because of the basic laws of physics.

Here's the deal: "User throughput rolls off the further you get from the base station," says Pulley. "[With] increased throughput, the signal becomes more sensitive to noise and interference. The further it has to travel, the weaker it gets."

So that means LTE cell sites need to be small and a have smaller radii than traditional macro sites to get the full data throughputs that LTE can offer, which will be up to three to four times higher than 3G HSDPA release 6, according to Adrian Scrase, CTO at the 3rd Generation Partnership Project (3GPP) .

"You've got to have new sites, and the capex implications of that are potentially horrible," says Pulley.

So future LTE femtos have to be cheap. PicoChip is developing a system-on-chip (SoC), code-named Feynman, that will enable a dualmode HSPA/LTE residential femtocell with a bill of materials of $70 in 2011. A lamp-post mounted, dualmode HSPA/LTE version of the femtocell will have a bill of materials of about $170.

"This is orders of magnitude cheaper than doing it any other way," says Pulley.

For equipment suppliers, though, the concept of small, cheap base stations, especially those that might supplant current base station models, is causing some tension, according to Pulley.

T-Mobile is keen on LTE femtocells. "Femtocells [will be] an important measure to supplement LTE for indoor coverage and capacity scenarios," says Frank Meywerk, senior vice president for radio networks at T-Mobile.
And China Mobile suggested that operators outside China could use the TD-LTE (Time Division LTE) version of LTE for capacity-enhancing femtocell deployments. That's an option for many European operators that, as part of their spectrum allocations for 3G UMTS services, have been awarded 5 MHz of time-division duplex (TDD) spectrum, along with their primary allocation of frequency-division duplex (FDD) spectrum.

TD-LTE would be a particularly useful choice for operators looking to maximize coverage in dense urban areas, as such deployments would "not interfere with FDD spectrum," according to Bill Huang, general manager at the China Mobile Research Institute. "I've heard that without having fixed spectrum allocated to femtos, it's not possible to deploy femtocells" because of interference issues.

Canada's Telus presented findings that femtocells do indeed have better performance in metro deployments.

"We're all here today to determine if there is a performance difference with going with very small cells," said Sam Luu, associate director of technology planning and strategy at Telus. "There is a significant performance difference at the edge and at the site."

But Luu cautioned that metro femtos are, for now, still only an interesting idea. "You can talk about technical better performance," he says. "But it still requires the R&D to get it off the ground. We're still in the early stages of evaluating the technology."

I have mentioned here and here that Femtocells can be considered as starting point for LTE rollout. Thinking about all the posts, it may be a better that an enhanced version of Femtocell or Femto++ is used. By this i mean that traditionally HSPA/LTE Femtocells are to allow max. 4 calls (more correct would be 4 radio links because the users could be in multi-call with CS and PS connection ... I know there is no CS for LTE but I am talking about HSPA) simultaneously and the Femto++ would allow 16 calls simultaneously.

In fact Huawei has a 16 channel Femtocell that is being trialled but not everyone is happy to refer to it as a Femtocell. A better term suggested is Picocell or my personal view is that depending on the power output, it could be classed as Femto++ or Picocell.

Operators can also reduce the cost of rolling out Metro Femtos by encouraging users to keep the access open on their Femtocells open and giving them a reward for every originating and terminating call made on their Femto.

More information on Metro Femtocells is available here:

No comments: