Wednesday, June 25, 2008

Does Anyone Care About WiMAX?

Apparently Intel swears by it, but does anyone else really care? We are already surrounded by several technology mediums that will allow us to stay connected: pre-4G LTE (UMTS Rev 8), 3G (UMTS), 2G, (EDGE, GPRS) and WiFi – do we need yet another technology (yawn)?

What advantages (if any) would WiMAX have over our existing technology base?

1) WiMAX is NOT more expensive to develop than LTE. In fact, with the recent creation of the Open Patent Alliance and the goal of capping WiMAX royalties at 10%, LTE deployments will be significantly more expensive due to higher costs of base stations, and device manufacturers will have much higher development costs (LTE is targetting 15% royalty model).

2) Intel commitment to WiMAX means that every device running an Intel chip will be capable of running on a WiMAX network . With this large potential device ecosystem, WiMAX has a very legitimate shot at being successful

3) For those who claim that LTE is an evolution from 3G while LTE is an alternative technology - that is just nonsense. Aside from some minor differences at the MAC and PHY layers, WiMAX and LTE are strikingly similar (SAE Gateway in LTE = CSN Gateway in WiMAX, PDN Gateway in LTE = ASN Gateway in WiMAX, both use OFDM transport, similar scheduling models, QoS models, etc). There will be no backwards compatibility for LTE to legacy 3G technologies - they are using different radio technologies (3G are CDMA or TDM, not OFDM), and therefore cannot be deployed on the same spectrum. This is why AT&T and Verizon have both heavily bid at 700MHz spectrum - for LTE deployment.

4) WiMAX is not really a replacement for T1 or DS3 Bandwidth. 3G is better suited to that. Quite frankly, if you are trying to replicate a deterministic bandwidth model (TDM) using a technology meant for burstiness (IP), you will spend a ton of money building a network that is built for SIGNIFICANTLY more capacity than it will ever see.

5) For those who are claiming that the span of time between production WiMAX networks and production LTE networks is shrinking - that is a very US-biased viewpoint. WiMAX networks exist today, LTE does not begin device certification until end of 2009. There are 3 issues that I point to that cause me to question LTE's time-to-availability:

a) With HSPA+ appearing, there will be significantly less incentives for European operators to migrate to LTE when they can leverage their existing spectrum, existing BTS, and existing handsets (backwards compatible) while still providing higher-speed data services.

b) With the telecom mentality that the LTE vendors have historically worked under, they develop for the next evolution of technology, not 3 evolutions down the road. So, with most worldwide operators deploying 3.6Mbps HSPA, there is still a roadmap to 7.2Mbps HSPA, 14.4Mbps HSPA, and finally HSPA+ (3GPP R7).

c) 3GPP R8 (official standards track of LTE) is slated for Stage 2 freeze in mid-2008 with official radification of Stage 2 planned for Dec 2008. WiMAX 802.16e was radified in 2005 and we still do not fully see an interoperability standard between vendors. This points to 2010-2011 before we will reasonably see anything similar in LTE.

Finally, anyone who claims that the move to 2.5GHz from the 900MHz and 1.9GHz of 3G means that the CAPEX model for 3G is significantly better than that of WiMAX - I think that is VERY highly imprecise. There is ALWAYS a tradeoff between capacity and coverage. If you want to attain the same coverage model for WiMAX as for 3G, you are right - you will need more towers. HOWEVER, if you are looking to create the same coverage AND increase capacity, regardless of the frequency used, you will need to create smaller cells with large overlap, meaning that the size of a 1900MHz cell providing high speed data services and the size of a 2.5GHz cell providing high speed data services will be roughly equivalent. You never get something for nothing.....

Labels: , , , , , ,

0 Comments:

Post a Comment

Links to this post:

Create a Link

<< Home