To print: Click here or Select File and then Print from your browser's menu

This story was printed from silicon.com, located at http://www.silicon.com/

Story URL: http://comment.silicon.com/0,39024711,11016838,00.htm

The Bloor Perspective: NT clusters, open source in healthcare and Bezos' brand new dilemma
This week Robin Bloor and his team of analysts take a look at Veritas' attempts to turn NT into a true high-end operating system; the upsurge of Linux in the healthcare sector; and the future of Amazon.com

By Bloor Research

Published: Monday 10 April 2000

There has been a well-recognised discrepancy between the actual availability delivered by popular platforms such as NT and the increasing dependency of business models on very high levels of uninterrupted service.

NT, and the hardware platforms on which it executes, has certainly continued to improve over time, but new pressures have kept business expectations accelerating away from the capability to deliver.

It is unfair to single out NT as being unready for 24*7. The plain fact is that most single-node server platforms will struggle to match the levels of availability that businesses now demand. We are not looking at reducing downtime by a factor of 50 per cent, but by several orders of magnitude. Simple clustering provides a partial solution, offering continued service through failover if the primary server node fails.

But this doesn't help address the related requirement of scalability. In order to provide scalability from commodity-class servers it is necessary to deploy a large number of nodes and to provide load balancing as well as failover.

This is essentially what Veritas has achieved with the Veritas Cluster Server for NT. Veritas has, for a number of years, provided this level of functionality for Sun Solaris and HP's HP/UX Unix systems. Bringing the technology to NT allows users to configure NT clusters of up to 32 nodes and to manage up to 256 clusters from a single workstation. Service groups can be defined within a cluster where each service group represents an application and the resources it requires (both hardware and software). Automatic or manual failover is supported to cater for unplanned or planned downtime for any particular server. The management workstation is just another application in this respect, and will also failover to an alternative node in the event that the node it is executing on fails.

While most of the industry is looking to towards high availability through special-purpose hardware, the ability to deploy commodity servers with a software-managed high availability potential will be very attractive to those mid-sized companies needing to migrate to a 24*7 operation. We also expect some interest from the ASP industry given the emphasis on service levels and scalability.

** Linux gets healthy **

In general, the healthcare market has followed the scientific market in its use of technology and thus platforms such as Mumps, which only saw niche usage in the commercial market, were used extensively in the health sector.

However, healthcare now suddenly finds itself leading the market, rather than bucking it, in its adoption of open source software and Linux. There are several reasons for this, not least of which is that funds for software development are often hard to come by in many areas of healthcare. However, there are also the facts that:

1. An open source implementation of medical record standards would speed their adoption and thus increase interoperability in a very critical area of health care.
2. The establishing of efficient health networks often runs into the problem of "Who pays?" and an open source approach to software reduces this problem.
3. An open source base which could be specialised by a variety of vendors may suit the needs of the diverse software users and budgets within the health care market.

There is a software company, the Minoru Development Corporation of Vancouver, Canada, which has positioned itself at the heart of the open source movement in healthcare, and which provides links to all the initiatives that are currently underway.

There are at least 18 open source projects in progress at the moment, which range all the way from a pharmacy research system to dental practice management software, and more are likely to start up.

In our view, this trend is a very interesting one. We expect to see open source take off far faster in this area that in any other simply because healthcare is a very strongly networked business and the commercial imperatives in this market are distinctly different from elsewhere. Additionally there are no strongly dominant commercial players in this market like there are in the CRM and ERP markets.

** Amazon tests the boundaries of brand **

Jeff Bezos, Amazon's chief executive, has been thinking about dispensing with the company's fixed assets to concentrate wholly on managing the brand. But can a company live by brand alone?

Internet companies are hard to pin down and may suffer a number of problems. Business may flow away from them in adverse circumstances. But sound brands are considered the only remedy to this problem. A well-established Internet brand is believed to have substantial value because it is a guarantee of customer interest.

The paradigm of brand value, and the example that excites Jeff Bezos, is Coca-Cola. The drinks company has achieved a huge capitalisation on the strength of its popular product. Distribution has long been contracted out to partners, while the company concentrated on advertising and promotion to keep Coca-Cola in the public eye. The strategy has been immensely successful over many years, although the share price has recently fallen from its best levels, along with many non-IT stocks.

But there is a big difference between Coca-Cola and Amazon. The drinks company is not just a brand - it is also a product. The drink formula is a closely guarded secret that is also protected by patents. For many years it proved difficult to create alternatives that tasted as good as Coke. So the brand was a mixture of a popular product and all the images built around it. A distributor of consumer products does not own a core product, and it remains to be proven that brand alone is sufficient to drive a major company.

At present, a factor in Amazon's success is its capability in handling the products it sells. In a sense, that could be seen as its core product which allies with the brand to create the company's value. If Amazon were to spin off its distribution skills, would the brand alone be enough to sustain the share price?

Of course, all this talk of the value of the Amazon brand comes at a time when there is concern about the ability of some Internet retailers to survive the downturn in investor confidence. Soft funds are no longer so readily available now that doubts have set in. Perhaps Jeff Bezos' talk of brands is to be taken as part of a PR campaign to maintain the Amazon share price.

For more analysis, see http://www.it-director.com