The Bloor Perspective: Intel and Linux; Orange and WAP; and Team Tandem make a comeback

Industry guru, Robin Bloor, and his team of analysts cast their eyes back over the latest top stories to hit the industry. Under the spotlight this week: the 'Lintel' alliance, Orange's WAP services and a start-up facing a rosy future

Intel has announced that it will now enable Gigabit Ethernet performance for Linux on Intel-based servers. It is providing a Linux Gigabit driver, an open source software driver, for its Intel PRO/1000 Gigabit Server Adapter. In conformance with the GPL (Gnu Public Licence), it is available free of charge to the Linux development community on Intel's Web site, allowing developers to customize the driver to meet their individual company needs. Linux is currently the fastest-growing operating system, and 90 per cent of Linux installations are located on Intel-based computers. But to be supported by Intel - a company that was once joined at the hip to Microsoft in the Wintel alliance - is a step change. A 'Lintel' alliance would make as much sense to Intel as a Wintel one. And if Linux takes off on the desktop it will make an awful lot more sense than a Wintel one. Software authors and ISVs may get worried about the implications of the open source movement, but chip vendors can rest easy. There is no such thing as a free chip, and there never will be. Orange dreams of a WAP Christmas UK mobile telephone operator, Orange, is set to launch a Wireless Application Protocol (WAP) service in November, ready for the Christmas rush. The initial services will be offered via a WAP-enabled phone from Nokia, and will be fairly basic, comprising information services including sport, news, weather and a business directory. However, given the general excitement about WAP, more sophisticated options will soon become available. The really interesting thing about the Orange service is the planned charging mechanism. WAP data calls will be charged at less than the price of a voice call in the first instance, but ultimately the plan is to cease charging for the connection and to start charging for the services themselves. Basic services will remain free while the premium content will be chargeable. This is fascinating stuff, as not only does it turn the existing mobile phone charging model on its head, but it also flies in the face of other, wire-based Internet services that already give away a great deal of what we might term "premium content." Consider share prices, for example. It used to be the norm to charge. The model then changed, and this information was given away by portal sites (such as Yahoo!) as a way of luring custom towards other services. Money could still be made from the share deal itself. Now we have services that do not even charge a fee for the share deal, as it is reckoned that hedging the funds and bulk dealing is worth far more than the 15 quid transaction fee. In other words, the tide is turning: whenever there is a higher level way of making money, the lower levels are given away. There is a caveat: as you (as a customer) move up through the levels, you need more information (and expertise). In the share dealing example, the free trading is only open to "experienced traders". So, what we have is two rules: Firstly, a service will be given for free as long as the supplier of the service stands to gain, indirectly, from the service. And secondly, if the customer needs help, then such help is also a service, which must be paid for unless covered by Rule 1. These two rules should come as no surprise as they are the twin pillars of retail pricing, understood and applied instinctively by customers and suppliers alike. Orange may intend to charge for premium WAP services - indeed, it's ideally placed to do so as it controls both the data feeds and the billing mechanisms - but when doing so they would do well to take these rules into account. The second coming of the Tandem men In April of this year, Austin-based Tantau Software announced the closing of $11m in an initial round of financing led by Austin Ventures, and also announced an agreement with Compaq. This involved acquiring the assets of Compaq's Tektonic Business Unit, and working with Compaq to provide a cross-platform application framework targeted at Web sites. The acquisition provided Tantau with an underlying customer base and allowed it to sell software from day one. Compaq holds about 8 per cent of Tantau and is partnering with the company by helping to provide worldwide support. Tantau's products include the Tantau Application Server Krypton - a component-based application environment with development tools InfoCharger - a high performance data analysis engine and Titanium, an inter-processor communications layer. All of these products are completely portable across Windows NT and various flavours of Unix. So what do they add up to? Actually, what they add up to, as near as damn it, is a portable, non-stop architecture. So will Tantau be successful? Clearly, it's hard to predict at this early stage, but its offering should be compelling if it comes as advertised. The product set constitutes a high-availability engine targeted specifically at very large Web sites and ISP operations. If it wins any of these high profile accounts and demonstrates that it can provide a non-stop service for them, then it will have customers queuing up at the door. Tantau also has an executive team that has proved itself to be portable. It includes Jimmy Treybig, the founder and former CEO of Tandem Computers Harald Sammer, a 22-year Tandem veteran John Sims, a VP with Tandem from 1985 to 1995 and Peter Robinson, a former Tandem sales manager in the UK. Silicon.com will be publishing an exclusive interview with Treybig on 11 November. * For more analysis, see http://www.it-director.com