We spoke with Manish Singh VP Product Line Management at Continuous Computing who provide the Trillium signalling protocol software for many femtocell vendors. He believes the case for LTE femtocells is compelling, that CDMA operators will be the first to deploy them and highlight that it’s not just the volume of mobile broadband data, but the type of usage which will drive network operators to offload data traffic onto femtocells in the near future.
What is Continuous Computing’s involvement with Femtocells?
We have a strong leadership position of protocol signalling software solutions, which we have tailored for femtocell applications, optimising for memory size and performance.
Our focus is on 3G UMTS femtocells including HSPA and TDS-CDMA derivatives. Going forward, we are seeing our LTE product offerings increase as we continue to unroll our LTE Trillium protocol offerings.
We have partnered with leading silicon vendors to integrate our software which will allow us to provide a basic platform for OEMs and we can also support other chipsets. Our stacks have been widely deployed and are “field hardened”. Our publicly announced key partnership for silicon is with picoChip.
We’re not doing so much on CDMA2000 yet – it’s a smaller market, and much of the demand we have received thus far has been for LTE products as it essentially allows for worldwide deployment.
Who do you think will rollout LTE first?
The first LTE deployments will be by the CDMA2000 operators, such as Verizon, Telus, and KDDI – they’ve all announced for LTE. Recently, the CDMA Development Group announced their support for LTE, essentially killing off UMB. Rev B CDMA will most likely be the last implemented release.
On the other hand, NTT DoCoMo too is leading the charge on LTE. Major UMTS operators have shown interest, and LTE will be the predominant 4G technology.
What do operators need LTE for?
The growth in wireless broadband data demand has been incredible. This is great news for the industry who need to increase “Feeds and Speeds” in their networks. LTE is primarily needed to meet the growing demands for wireless broadband data.
LTE will initially be used only for data traffic, so the first consumer devices will be data dongles – voice capable LTE handsets will follow later. Long term, operators will carry both voice and data over LTE’s converged packet core.
What is the role of femtocells in LTE deployment?
Femtocells are more important in the LTE rollout than for HSPA data.
CAPEX investment in nationwide network rollout is huge. Femtocells offer a new rollout strategy, providing one femtocell, for one consumer, at a time. This allows operators to scale their rollout in a controllable fashion, aligned with CAPEX investment.
Also, remember it does take time to rollout and achieve mass market adoption of new wireless technologies – 3G UMTS was due for commercial launch in 2002 which only became mainstream in 2006 or 2007. This almost 5-year delay pushed mobile operators to realise any ROI on 3G. LTE should be commercially available from 2010 at the earliest. LTE Femtocells allow operators to scale their CAPEX investment as adoption scales.
Spectrum is another key reason that tips the equation in LTE femtocell’s favor. For LTE there are three main options:
- 2.6GHz, which seems popular with European regulators
- Re-farming the 900 MHz GSM band
- Digital Dividend (redeploying the analogue TV spectrum)
2.6GHz offers the biggest challenge because the higher frequency will not penetrate in-building or achieve such long distances. Indoor coverage will be a bigger problem in 2.6GHz. On the other hand, re-farming 900MHz requires nationwide coverage to meet the regulatory requirements and that means huge upfront CAPEX investment.
There is not a more cost effective way of providing the capacity than by using femtocells:
1) It cuts down on OPEX by transferring the operating costs to the consumer (power, site rental, backhaul etc.)
2) The traffic mix and kind of data being used is changing. It’s not more bursty in nature and so needs dedicated localised capacity.
Finally, an important point here is the speed of domestic wireline broadband that is needed for backhaul. LTE femtocells will require a bigger pipe to/from the home.
Won’t the reduced speeds of wired broadband limit uptake of LTE femtocells in the short term?
Verizon and ATT are already deploying FTTx which will satisfy the requirements for LTE femtocells. This gives benefits to operators with both fixed and wireless assets.
The rapid uptake of IPTV (Telco TV) is already creating the required backhaul infrastructure for LTE femtocells. Piggybacking LTE traffic over their existing FTTx infrastructure allows operators to provide increased wireless data rates and capacity indoors, where it is needed most. This also accelerates operator return on fibre rollout investments.
The FTTx infrastructure provides for:
- “Inside-Out” rollout of femtocells [Deploying LTE femtocells first with fewer outdoor sites]
- Scalable business model
What are the technical and commercial issues where femtocells and wireline broadband are provided by different service providers?
Today, wireline broadband operators already have the capability to differentiate and provide managed QoS for 3rd party services. If not, they could easily implement this using Deep Packet Inspection (DPI), which analyses and shapes IP traffic.
Commercial arrangements are much more important. Without these, you can use best effort, but risk poor QoS and a poor or inconsistent customer experience.
For example, a home may be sharing a wireline broadband connection between a web surfing session, a video download, and a voice call on a femtocell. If the voice call isn’t prioritised then it could cause problems, i.e., a dropped call or broken connection.
So why should ISP’s prioritise femtocell traffic?
It all comes down to the commercial arrangements. Some sort of revenue sharing might be needed for ISPs to prioritise femtocell traffic.
Also, note Comcast has already announced wireless services over cable and Cox plans to deploy it’s own wireless network. ISPs too can provide their own wireless services with femtocells.
To prioritise traffic, operators can use Deep Packet Inspection and commercially offer tiered SLAs. Premium QoS services can be priced at a premium.
Why would wireless operators need Deep Packet Inspection (DPI)?
DPI is actually more compelling in wireless networks than wireline. There is a tsunami of data traffic coming for wireless operators with future growth predicted to be off the chart.
Wireline operators can throw fibre at the problem, but wireless operators are constrained by limited spectrum.
The 3G HSPA system was designed with a shared channel which delivers peak rates to individual users one at a time. It was designed for short, bursting data traffic. The game changer is that the type of data traffic that uses mobile broadband is shifting. Mobile video technologies, such as YouTube, are becoming more popular and grab the network bandwidth for longer periods.
When wireline broadband was launched, peer-to-peer, file sharing traffic quickly consumed up to 80% of available capacity and hogged the bandwidth. Wireless operators need to adopt a traffic shaping approach using DPI to manage their available capacity. This will lead to more satisfied customers and maximise network usage.
DPI also plays a key role in securing the IP network. There are frequent DoS (Denial of Service) and Zero Day attacks which require adequate security detection mechanisms. Up until now, DPI was typically deployed as a “bump in the wire” – an extra box placed in the network path that all traffic passes through. We believe it makes more sense in the long term to incorporate the DPI into the gateways themselves, thus reducing latency.
How do you see the migration to an All-IP wireless network?
Most operators have a long term vision for a single core network which supports both voice and data.
Their first priority should be to move all packet data traffic to LTE and an evolved packet core.
Later they can migrate the voice traffic, which will reduce their Opex and simplify their networks.
The challenge is no different between migrating voice to an IP core and migrating femtocells to LTE. There are many additional considerations including E911 emergency calling, traffic interception, and voice supplementary services which need additional consideration.