Taking the pulse of HP-UX

The local availability of the latest version of Hewlett-Packard's proprietary implementation of the Unix operating system immediately raised the question: is it still relevant over 20 years after it was introduced?

With the rise of Linux, is any flavour of Unix still a significant player in the server market?

Apparently it is. At least according to HP's Andrew Fletcher, who manages the company's business-critical servers unit in SA. He says from a revenue point of view, HP Unix (HP-UX) has been a top player in the local Unix market for a long time.

Is this overkill or simply HP getting real about future storage requirements in the information age?

Warwick Ashford, portals managing editor

Fletcher says HP-UX has a more or less equal share of the market as competing operating systems from Sun Microsystems and IBM. Although unwilling to commit to a number, he said there were several hundred customers using HP-UX in SA.

Obviously for those HP customers a new version of HP-UX is bound to be useful, but surely newcomers to the market will opt for something like Linux?

Again, Fletcher disagrees. He backs up the official HP line of lower total cost of ownership (TCO) and better virtualisation and dynamic resource management by saying that in the past year, HP has proven lower TCO for customers needing these features. He maintains that when it comes to maturity in virtualisation, Unix beat Linux.

Fletcher asserts virtualisation is a key strength that puts Unix ahead of Linux, but says HP-UX is the clear leader, claiming HP's suite of tools to be the most comprehensive. He also says HP has put a lot of work into reducing cost and maintenance to make it competitive.

Is it possible that HP-UX's age is really a strength and not a weakness?

If we accept that HP-UX has had a much longer time to perfect things like virtualisation and workload management through dynamic resource allocation, what about performance?

HP says HP-UX 11i version 3 scores well here too. The company claims an average 30% performance increase in applications with the new version.

How's this possible?

It all comes down to careful work on the operating system's kernel, says Fletcher. Apparently this has been a major focus of version 3, which has involved some rewriting of the code to clean it up, as well as optimising it to work more efficiently and adding new virtualisation capabilities.

HP says these new features will enable administrators to move memory and resources among distributed virtual partitions on the fly with no disruption. With the performance gains HP is claiming, one can only wonder why HP has not tinkered with the kernel in this way a bit sooner.

But so much for performance, what about HP-UX's ability to deal with rapidly increasing storage demands?

HP claims it has this angle thoroughly covered. HP-UX has followed Sun's Solaris 10 into zettabyte territory. Sun predicted others would follow, but when introduced its zettabyte file system for Solaris, dissenters described it as a feature no one needed. HP obviously disagrees.

Fletcher says data warehouses are averaging 100 terabytes and storage systems are already been measured in petabytes (a thousand terabytes), so the next logical step is zettabytes (a thousand million terabytes).

Although zettabytes may be the next logical step, HP-UX 11i version 3 could be said to be taking things a little too far. From a capacity of just 32 terabytes in version 2, version three is capable of addressing not one, but 100 million zettabytes of storage.

Is this overkill or simply HP getting real about future storage requirements in the information age?

As always, only time will tell if HP's plan to continue investment in HP-UX is well founded, but for now Fletcher is confident the operating system has a future. He says indications are the demand for storage will continue to increase, and HP-UX will be ready.

Investing in effectively limitless data storage sounds like a good bet to me. What do you think?