HP OpenView - why users daren't speak the truth

Peter Judge: Users are dumping HP's OpenView management system, so why are they scared to talk about it?

For some time now, network management companies have been universal in their deprecation of what they call the "legacy" systems, HP OpenView, IBM's Tivoli and Computer Associates' Unicenter. But their scorn is most of all reserved for HP OpenView.

It's old, I am regularly told. It is expensive, and that cost is compounded by the fact that it is so labour intensive. "It can't do Layer 2 discovery," I am frequently told, meaning that it can't tell exactly what is on the network, so network administrators have to spend their time redrawing network maps to make them reflect reality, rather than doing actual useful work with the management product they have bought.

A new day is dawning -- upstart network management companies have been saying for the last five years or so -- and legacy management products are scheduled for the compost heap of history.

Interesting story, but to convince me that this is true, you would think these companies could line up customers to back the story up. "Let me talk to someone who has dumped OpenView in favour of your product," I ask."We'll get back to you," they say. And usually they don't.

It's not that people like the legacy stuff. I've had even more difficulty trying to find a satisfied HP OpenView customer who will come forward and say they prefer the product to its newer rivals.

So I have come to the conclusion that most people who manage networks are miserable souls, still reluctantly tied into legacy network management systems. They want to get rid of them, they are using the alternatives where possible, but they haven't been able to dump OpenView or its chums.

On that evidence, it looks like the big revolution in network management is just not happening.

So, I was very pleased when a small, up-and-coming network management company promised to line up a major customer, a large IT services firm, who would dish the dirt on just why it was dumping a major legacy network management system. Here at last was some evidence that the big shift in network management that I'd been hearing of might be happening for real.

Some weeks later, on a grey afternoon in London, I dialled a bridge line and started to introduce myself to a group of trans-Atlantic execs from the management company and its customer.

The customer has a great story. It handles outsourcing for customers, giving it a total of more than a million seats, and it manages all this from 85 network management consoles. The firm made the decision to remove OpenView from their network in October 2001, and it is now 90 percent there.

Since the company takes on and runs customers' networks, every new customer means another tranche of OpenView stations to replace. One recent customer came to them with around 100 "legacy" (ie OpenView) management consoles -- that same network, outsourced, is managed by only three of the new consoles.

Will there be a remnant of OpenView that can't be taken out? No, said the user, in fact the rate of replacement gets faster. "As we refine our implementation of the new product, it becomes more flexible and handles more exceptions. The pace is accelerating, and the end point is very soon."

Apparently, the enthusiasm for getting rid of OpenView even extends to analysts with a long career in the product. "We anticipated this might be difficult, because they have spent a long while writing code around the legacy system," said the spokesperson. "In fact, they were delighted to leave that behind, get more creative, and start working at a higher level."

With the new system, these analysts have begun to work on managing optical equipment, resolving problems much more quickly, and getting new customer networks up and running more quickly. "Customers come to us using six to ten different management tools," said the user. "We come in lean and mean." The new management product paid for itself within 12 months.

All this is interesting, but early on in the conversation, we hit a snag. "When you write this up," says the customer, "we'd rather you didn't mention the name OpenView."

My heart sinks. I explain to them that this somewhat detracts from the purpose of the article. Correct me if I'm wrong, readers. Maybe I'm losing my news sense, but I feel sure that you will not be all that interested in a story headed "Company rejects un-named management product in favour of one from a small company that you probably haven't heard of."

But the customer is adamant. Even when the OpenView consoles are completely gone, my friends there would still rather not rock the boat, since they have other relationships with HP. "We buy hardware from them," said the spokesperson. "It isn't politically correct to come out and trounce them."

"It's best not to throw stones at our neighbours, since we might end up working for them." said the user. "We have to team up in many areas with our fiercest competitors." Right, fair point. Only you're obviously not competing that fiercely if you won't mention them by name.

Myself, I don't see why HP's strength in hardware should keep the public from knowing the low opinion users have of its (still apparently) market-leading network management software.

So, I've taken the only other obvious course. Here is the story, with the name OpenView included, but without the name of the customer. I'm sorry, guys, it sounds like you have done a great job, and I'll happily tip my hat to you in public when you're ready to put your name to the full achievement.

To have your say online click on TalkBack and go to the ZDNet UK forums.