The Bloor Perspective: Biometric choices, 3G+WLAN and EMC buying VMWare

Biometrics have been used for many years for a variety of applications across a range of industries, including the use of fingerprint or iris scans for personal verification in the financial services industry, such as for withdrawals at ATMs. But much of the evidence for or against biometrics can be found by looking at the travel industry - specifically the measures being taken at airports across Europe.

Biometric identification has been in use at airports in Europe, and even worldwide, for a number of years but it is in the past three to four years that implementations have been accelerating. Schemes in place in 2001 include a facial recognition programme at Keflavik Airport in Iceland, as well as the start of an iris recognition programme for frequent travellers using Schiphol Airport in Amsterdam. Yet it was the terrorist events of 2001 that provided the greatest impetus for airports to jump on the bandwagon and initiate their own schemes.

The biometric schemes in place across Europe are a mixed bag: some employ just one of facial, fingerprint and iris recognition; others - such as those being implemented at Frankfurt Airport in Germany and Umeå Airport in northern Sweden - combine both fingerprint and iris recognition capabilities. In addition, some airports are using biometric identifiers to control access to restricted areas at airports, while others are aimed at speeding up the processing of passengers. Further, some schemes are being extended to cover both employee access as well as passenger security and convenience.

As part of its efforts to improve homeland security, the US is mandating that biometrics must be incorporated on the passports, visas or other identification documents being used by passengers entering or leaving the country under its US-VISIT scheme. However, the number of schemes - as well as the variety of different biometric identifiers in use in these schemes - will make interoperability problematic, especially if machines have to be put in place to automate the identification of individuals through fingerprint, face and iris analysis.

The International Civil Aviation Organization is working to develop a global, harmonised blueprint for the integration of biometric identification into passports and other forms of identification. It is hoping to implement a standardised system of identity management among its 188 member countries. It has selected face recognition as the preferred biometric identifier owing to this form of biometric identifier having the highest compatibility with key operational considerations, followed by fingerprint and then iris recognition.

However, these findings fly in the face of trials of face recognition technology in various parts of the world, which show that false acceptance rates are high using this technology and that it is not especially effective in matching people against databases of known undesirables.

Fingerprint recognition, on the other hand, is more reliable since a computer can match up to 70 different characteristics that are unique to an individual. Iris scans are even more effective, able to match up to 270 unique characteristics.

Within Europe, there is a wide variation of biometric techniques being used, meaning that the desired goal of standardisation is unlikely to happen any time soon. What is required is that the internal ministers of the EU countries get together to reach some kind of compromise to promote interoperability but this is unlikely to happen any time soon, especially in the light of the expansion of the EU.

In the meantime, many schemes involve the use of more than one biometric identifier but this makes such schemes more expensive to roll out. The UK government, for example, is toying with the idea of including face, fingerprint and iris biometric identifiers on its proposed ID cards. For many concerned with privacy, that is going just too far. Question: Just how many radio bands can a mobile phone support? Answer: As many as the user will pay for.

Over recent years, phones have become dual band, tri-band and even quad-band to cover an ever larger frequency range of countries with GSM networks. Even newer 3G phones have the combination of Universal Mobile Telephony System (UMTS) with GSM bands. Many also have Bluetooth, so why not add wireless Local Area Network (LAN) to 3G?

The combination of licensed and unlicensed spectrum might seem like maverick demonstrators working alongside government officers but it will generate some interesting opportunities.

Of course there are technical challenges to overcome, the most important of which is seamless hand over from a wireless LAN, or Wi-Fi, network to the 3G cellular base station when moving out of range. Those who have tried early 3G to GSM hand over will already know this is by no means an easy problem to fix. What will happen to roaming and billing, especially in a predominantly pre-pay world? Then there is the potential problem of frequency interference with Bluetooth.

But the business potential is broadening, as are those expressing interest. NTT DoCoMo started trials of combined handsets earlier in the summer and is now rumoured to have completed them. They reportedly have met with some success but DoCoMo is not overly bullish and states that more work needs to be done.

One company driving the combination of technologies forward with some success is Florida based Calypso Wireless. It claims its proprietary technology provides switching between cellular networks and wireless LAN without losing connection. When a suitably equipped mobile phone detects a wireless LAN it automatically switches over but remains connected to the cellular network until the new link is established. Calypso works with established handset manufacturers to provide suitable cellular phones but its bigger challenge will be winning over mobile operators.

Why is wireless LAN versus 3G such a big deal? Of course there is the pure speed argument. Today's mobile phones manage data rates of around 9.6Kbps, perhaps as much as 40Kbps. 3G networks will probably be able to reach 385Kbps over the coming years. Existing wireless LAN technology provides 11,000Kbps and can already go faster.

Then there's the market opportunity this opens up for new operators. Cable companies, fixed line operators, even hotspot providers can all pick up a piece of the mobile data pie. That would be bad news for mobile operators, especially those with expensive 3G licences, although progressive mobile operators might consider this an additional opportunity for them to move into the wireless LAN and enterprise market.

But there is a big caveat. What's the killer app and who will pay? Mobile email only requires relatively low bandwidth and although home broadband raises user expectations about web surfing, and mobile video has had a lot of advertising push, there's a limit to how much an average user will pay.

Corporate usage is different. Wireless LAN pricing models that ignore megabytes but measure connection time will be more appealing than cellular tariffs that price by the data volume transmitted.

Applications such as mobile video conferencing might at last take off! Combining 3G and wireless LAN on mobile handsets could be the start of something big, changing further the way we all work. EMC's acquisition of VMWare is a truly bold and surprising move. VMWare has been the much sought after target of many potential acquirers. In case you are not aware, VMWare provides a virtualisation capability for x86 chip architectures (meaning chips from both Intel and AMD but not Itanium as yet). The value proposition is 'utility computing' - the ability to set up multiple virtual machines within a single server. VMWare has some extremely positive case stories in this respect and has built up a head of steam in the market place both in the US and Europe.

Microsoft tried to negotiate an acquisition of VMWare a while ago but VMWare was not easy to acquire - it was not a VC financed operation, so the founders had complete control over the destiny of the company. Similarly there have been rumours that both IBM and HP were interested in the company. Thus it is a considerable coup for EMC to have won this prize. The founders of VMWare must also have been happy with the $635m deal and not just because of the pay-off.

VMWare was expected to move to IPO soon and the shareholders would have had ample opportunity to enrich themselves if they had wanted to pursue that route.

Is there any synergy between EMC and VMWare? There is some. For VMWare to operate with maximum effect on servers (it works on the desktop too) it really needs to sit over a SAN architecture. Technically, it is possible to move applications about from server to server (and VMWare can do this in a reliable way) but it is not possible to relocate data dynamically. You need a SAN.

However, the acquisition of VMWare has to be seen in the light of EMC's other recent acquisitions: Documentum and Legato. EMC is clearly building a software business - and doing it by acquiring what look to be very valuable assets. There can be little doubt that it will be building a services arm to exploit these too.

EMC is beset with the problem that many successful hardware companies run into - it had saturated its primary market and needed to diversify or run the risk of stagnating and possibly declining as other vendors attack its prime market.

EMC is pursuing an imaginative acquisition strategy. It is not an easy thing to pull off but for now the signs are good. It will be interesting to see how well it does.