US operator Cingular confirms the 'clues'
Close
By Mike Woolfrey, Research Analyst (from GSM'centric EMC)
Cingular Wireless, the USA's second largest carrier, has finally confirmed its technology migration path. The company announced its decision at a press conference in New York (30 October 2001), with the CEO Stephen Carter and CTO Bill Clift outlining the basis of its choice.
In opting to overlay its existing TDMA/AMPS footprint with GSM/GPRS/EDGE, Cingular is confirming the strategy that many (including EMC ) had foreseen. Stephen Carter described the decision as one of 'best known secrets' in corporate America. The company had left what it described as a number of clues throughout 2001. These included the decision to strip CDMA out of Seattle and replace it with GSM and the recent network sharing deal with the GSM operator VoiceStream (). What was slightly surprising was the heavy dependency on EDGE and GAIT and that UMTS/W-CDMA is not currently part of Cingular's plans.
Source: Cingular Wireless and EMC World Cellular Database Figures include Puerto Rico
What's the deal?
Cingular is to deploy the services of three vendors (Ericsson, Nokia, Siemens) to overlay its TDMA footprint. TDMA accounts for 70% of the company's POPS, the remainder are served by Cingular's GSM-1900 systems. Each of the three will supply radios, switches and handsets. Nortel networks recently signed a contract to supply Cingular's backbone network (), but would no doubt have been disappointed not to pick up any of the radio work. Cingular also recently announced a supply deal with Lucent (). This deal was ambiguous, prompting speculation that Cingular may well be looking at CDMA again. In hindsight this looks as if it is merely a deal to maintain the existing TDMA network, whilst the others go about the network overhaul. Lucent's decision to cease GSM-8/19 production eliminated any chance of it benefiting from Cingular's migration.
No specific details were given to the nature or split of the supply deals. A total of $3 billion is to be spent on the network migration and capacity improvements. Cingular was keen to stress that a proportion of this money would have to be spent anyway, and the $3 billion is not all on overlay. The overlay equates to $18-19 per POP. In choosing Siemens and Nokia it is the first time Cingular or its parents (the former BellSouth and SBC) have given a supply contract to either. Although Siemens is a large player in GSM infrastructure, it has little or no contract history in the USA. Earlier in 2001 it signed a handset supply deal with Cingular, its first in the USA (). It is possible that Siemens has the smallest share of the contract. Nokia has a limited infrastructure history in the USA, supplying Aerial/VoiceStream and more recently AT&T Wireless. It does however supply both TDMA and GSM handsets to Cingular, amongst others. Ericsson has a far greater history with Cingular and its parent companies, having supplied BellSouth with AMPS and TDMA equipment since 1986 and Southwestern Bell since 1984. Ericsson also supplies Cingular's (PacBell) GSM/GPRS network in California and Nevada.
Cingular is currently offering GPRS services in a number of its former BellSouth Mobility DCS and PacBell Wireless markets. All current GSM markets are expected to be offering GPRS services by Q1 2002, with these same markets being EDGE ready by year end 2001. Handsets will follow later.
2002 Deployment of GSM/GPRS in current 850MHz 2003 EDGE deployment largely completed 2004 100% of markets EDGE-enabled.
With vendors now supplying EDGE-ready base stations, it is likely that Cingular will meet its network deployment targets. EDGE is a relatively easy migration from GPRS, requiring a remote software upload to the base station. The key to the networks being ready on time and the customer migration from AMPS/TDMA to GSM comes in the handsets.
Cingular had been lined up to make this announcement on a couple of occasions earlier in 2001: a cancelled press conference at the 2001 CTIA show was believed to be a possible announcement date. Cingular was waiting to be sure that its vendors would offer GSM in the 850MHz band. Unlike AT&T Wireless, Cingular does not have a great deal of 1900MHz spectrum. When opting to overlay with GSM it had to use its 850 properties, spectrum that is already being used for AMPS/TDMA networks. In doing this, it would become the first operator to do so, plus it would have to make any changes whilst the network was still live. In markets such as Chicago where Cingular has low spectrum holdings (25MHz) dedicating spectrum to what will initially be a traffic free network will not be easy.
In opting for EDGE, Cingular has opened itself up for a certain amount of scrutiny and, in many quarters, criticism. To date Cingular is the only carrier to depend so heavily on EDGE. Whilst AT&T Wireless has made noises that EDGE is part of its plans, it is far more likely (given its relatively large 1900MHz holdings) to bypass EDGE in most markets and go straight to UMTS.
The whole role and performance of EDGE has been heavily debated within the industry ( and ). Nokia has made claims that it has seen data rates as high as 473Kbps, but more realistic average throughput rates could range from 60-160Kbps. At Cingular's press conference Ericsson showed EDGE running at 160Kbps. In theory EDGE will run at three times the rate of GPRS, so with some networks only showing data rates of 15-20Kbps, the lower end of the quoted range maybe more applicable. Cingular was keen to stress that the performance of EDGE is not fully appreciated: whilst it does not bring an immediate voice capacity uplift, techniques such as adaptive rate vocoding and antenna and frequency hopping will bring voice capacity improvements. EDGE will also act to improve battery life when compared to GPRS as it uses fewer timeslots for similar data performance.
With EDGE still only really in the lab testing stage, it is an unknown element. There is scepticism about whether it can perform in a live network situation, with performance at the cell perimeter a possible hindrance. However, one would assume Cingular must have seen some sort evidence to make it trust EDGE so heavily.
Cingular's plans make no timescale reference to UMTS. This is largely due to its spectrum holdings. For it to deploy UMTS it would need 5MHz of unused spectrum, and with an average of around 25MHz per market, this is something it does not have. The FCC's move to eliminate the spectrum cap will no doubt help facilitate any move to UMTS, to the extent that Cingular is likely to try and acquire the smaller operators in markets where capacity is tight. This will give it two options; one is to move existing traffic (850 and 1900MHz) onto any PCS spectrum it could acquire, and then clear the 850 band and the other would be to go straight to UMTS in the 1900MHz band. In markets where Cingular only has PCS spectrum (LA and San Francisco) it is extremely unlikely that it will pick up any 850MHz spectrum. Whilst UMTS is certainly on Cingular's long-term roadmap, it sees its current spectrum holdings as sufficient to last another five years. To that extent, UMTS for Cingular looks unlikely before 2006.
Handsets
Cingular Wireless' plan is based on two new handset products. Firstly, Cingular is opting to build the majority of the GSM network at 850MHz. As yet there are no single mode GSM-850 or dual-band GSM-8/19 handsets on the market. In fact none of the three handset suppliers have yet to release details of proposed devices. The second key point for handsets is GAIT. GAIT (GSM ANSI-136 Interoperability Team) handsets are dual-mode (in some cases tri-mode, with AMPS) that will allow users to switch between GSM and TDMA coverage.
Until now GAIT handsets had been seen as some what of a dream (). Only Siemens had made any sort of product announcement, with Nokia and Ericsson tight lipped. If the commitments made by Siemens, Nokia and Ericsson are met, then it appears that GAIT may well come to fruition. Although little precise detail (eg terminal number, specification) was given, Cingular is to begin selling GAIT handsets in March 2002. These will initially be TDMA-8/19 and GSM-1900, but models using the 850MHz band for GSM will be available from July 2002. Models will be GPRS ready, but it is unlikely that GAIT handsets will feature EDGE. GAIT handsets are targeted as a transitional product, and it is likely that EDGE availability may coincide with the completion of the GSM overlay.
The exact role of AMPS in any GAIT handsets is unclear. AMPS has provided large areas of footprint for rural areas of Cingular's network. TDMA and AMPS have coexisted in the same handset, allowing customers to switch between the two networks. Cingular continues to migrate its AMPS customers across to digital (18% of its base were AMPS at September 2001). AMPS does however provide lucrative and strategic roaming coverage.
Source: EMC World Cellular Database 2001 figure is for June
Both Ericsson and Nokia have indicated that some of its GAIT handsets will have AMPS functionality. It is unclear if Siemens will have the same, but its S47 will not operate on AMPS networks.
Is this the right move?
Whilst Cingular has been evaluating its various network options, the industry has been pushing the various options it had. These were essentially two: one was to do as it has, the second would have been to install a totally new radio system in CDMA. Following AT&T Wireless' move, it is widely accepted that TDMA operators will need to migrate to another RAN technology to offer future services (3G). Rightly CDMA companies have seen this as a big opportunity, especially when TDMA operators, CANTV in Venezuela for example, are confirming that CDMA will be replacing existing TDMA/AMPS systems. The logic behind this is clear: whilst a new RAN would be required, the existing ANSI-41 core network could be retained. The performance of CDMA over GPRS/EDGE was also pushed as a reason to move to cdma2000 1x. Whilst EDGE may bring some increase in voice capacity, it is not in the region of 50-100% (from existing cdmaOne systems) that cdma2000 1x will bring. Data rates for EDGE are perhaps comparable (although yet to be proven) with cdma2000 1xRTT. But with the next phase to cdma2000 1x (EV-DO) only requiring a new channel card and a software load, compared to UMTS (if Cingular gets the spectrum), which requires virtually a complete new network, Cingular had the option to make major network changes now, but only once. If it were to go to UMTS, wholesale changes would be required again.
Cingular has undoubtedly looked at this, and by evidence of its decision rejected it. By opting to go for GSM it hopes to gain the associated economies of scale. Whether or not these economies realise themselves remains to be seen. An increasingly large number of elements of 2.5/3G networks are common to the various radio technologies (HLR, prepaid platform, PDSN, WAP gateway) and with protocol changes switches can be transferred. Added to this, elements of the RAN are transferable. This includes physical elements such as cabinets and racks. So one could question just how much will be saved through the specific radio elements, especially when GSM-8/19 equipment is being deployed, and will be deployed by a considerably smaller market than GSM-9/18.
Probably the biggest savings can come from GSM handsets. GSM handsets have been historically cheaper than CDMA/TDMA, although parity between cdma2000 and UMTS looks more likely. This is not necessarily going to be the case for Cingular. In using GAIT/EDGE/GSM-8/19 handsets it is going to be fairly unique. Admittedly, other TDMA operators will follow suit in offering GSM-850 band and AT&T Wireless will almost certainly look to put GSM in at 850. But with all things taken into consideration it is still only a small market. GAIT handsets are going to be more expensive than existing TDMA or GSM offerings, plus choice is going to be limited. So for Cingular to stay competitive at the consumer level it will have to subsidise handsets. This is something it is willing to do, indicating that the additional subsidies are worth the protection of its subscriber base, and GAIT handsets are needed to prevent a massive rise in churn. Added to this, the extra expense of GAIT handsets may have been more cost efficient than roaming on competitors' (AT&T Wireless and VoiceStream) GSM networks, which would have been needed to keep a nationwide footprint.
EDGE handsets are yet to emerge, and with European operators likely to bypass EDGE and go straight to UMTS, possible markets are limited. The vendors point out that GPRS handsets will evolve to EDGE. But with GSM covering four bands (850, 900, 1800 and 1900) are manufacturers going to produce quad-band phones, for what is likely to be a limited market? This does not even take into account that UMTS will have to be factored into them at some point.
What does this mean to the US industry?
Cingular has made a very awkward decision. Its spectrum position is making it even harder, but in opting to go for EDGE it has only partly solved the problem. Whilst UMTS is not in the migration plan, it still looks like it is on the roadmap at some point. The economies of scale may or may not bear fruit, and with $3 billion to be spent, further expenditure will be required to move to UMTS.
CDMA suppliers Lucent and Nortel along with Qualcomm will be disappointed that Cingular has taken the GSM route. Lucent was believed to have campaigned hard for Cingular to overlay with CDMA. In the long run the CDMA camp may well benefit from Cingular's path. 2002 is set to be a very competitive year, with both of the CDMA operators (Sprint and Verizon) offering 1x services and fellow GSM operators (VoiceStream and AT&T Wireless) with pretty much nationwide GPRS services. This could put Cingular at a large disadvantage, as its initial GPRS coverage area will be less than that of its competitors. Its handset choice is likely to be more limited than others and potentially more expensive, especially if any sort of roaming is required. Therefore Cingular could find it hard to compete in both the voice and data markets. Limited spectrum and the requirement for cutting over new networks on top of live spectrum will stretch capacity, making large voice bundles for less dollars harder to support.
In Cingular's defence, it has managed to build up the second largest customer base in the USA by operating largely regional coverage. Prior to its creation, SBC and BellSouth marketed through 13 separate brands. The pie chart below also shows that since its creation in Q4 2000 Cingular is ranked third in terms of net adds amongst the national players. Whether this can continue into 2002 is open to examination.
Source: EMC World Cellular Database Market share is based on the net adds of the six national operators, and excludes regional operators Figures include the subscriber restatements by Verizon and Cingular
The partners
There has been a certain degree of speculation that BellSouth may exit the joint venture following Cingular's decision. Part of the delay in the strategy announcement is thought to have been created through BellSouth's opposition to the GSM-based path, preferring a cdma2000 migration. BellSouth's TDMA networks in Latin America are believed to be moving to CDMA, and although this is yet to be confirmed. This further points to its preference for a CDMA path. If BellSouth was to opt to sell its 40% holding in Cingular it would be a big change of direction, as the new venture was only formed in October 2000.
If BellSouth was to put up its share for sale, it may well look to join Sprint PCS. Purchasers could come from those looking to establish global brands, including the possibility that France Telecom could try to bring the Orange brand to the USA. However, with the economic climate none too healthy in the telecoms world, financing could shelve this issue, for the foreseeable future at least.
e-searchwireless.com
----------
"..However, one would assume Cingular must have seen some sort evidence to make it trust EDGE so heavily."
With its hatred of Q and cdma, Cingular had no choice - with or without evidence. |
