NetEvents Portugal 2008

Last week saw yet another successful NetEvents event, this time in sunny Portugal in the Algarve.

A wide range of vendors attended, and the event kicked off with a keynote by Bernd Schumacher from Nokia Siemens Networks. He argued from both business and technical standpoints that the only way to connect five billion extra people to the Internet was to use IP-based Carrier Ethernet.

While no-one seemed likely to disagree with his argument from a technical point of view, the problem for carriers is, as ever, how they're going to get from here to there.

Right now their networks are full of old and often not-so-old TDM-based equipment. This is stuff that supports circuit-switched voice traffic and is perceived in the industry as being outdated. You'll find the work 'legacy' applied pejoratively to this technology - mainly by vendors of the new stuff.

This means though that there's a huge cost sunk into TDM equipment. To rip and replace would be hideously expensive -- and even if the world economy were as robust as it was two years ago, no carrier can afford simply to chuck it all away and start again. So Schumacher argued that carriers need to build business models that allow them to develop and sell services, thus generating revenues to fund new networks.

The problem is, as Schumacher did acknowledge, that his telco customers have not traditionally been that successful at doing this. Instead, third parties have come along and built services on top of their networks, and so creamed off the revenues. Think Yahoo, Google and YouTube.

Following a light grilling on stage from the moderator and the audience, Schumacher made way for the first plenary debate session. This was on the CIO's perspective of enterprise networking, introduced by Evelien Wiggers from IDC. Along with the panellists, the discussion covered how CIOs can manage the increasingly complex networks of which they are in charge. This includes issues such as device management, which is showing signs of getting out of control as suers continue to attach phones, PDAs and memory sticks to the network.

Debate II asked if the mobile operators have done enough with their recent femtocell initiatives to see off the challenge from Wi-Fi. Introduced by Dean Bubley of Disruptive Analaysis, the panel argued one way and the other, with the wireless vendors such as Trapeze pitching for Wi-Fi, while the likes of Motorola argued that the operators needed to improve coverage - via femtocells.

With Debate III, the event attacked the issue of interconnecting islands of Ethernet across the globe. The question here is how carriers can link up their differently configured and specified Ethernet networks to deliver a seamless and easily managed experience for the enterprise. The debate, under the command of Gartner's Ian Keene, concluded that the key issue was manageability.

The final session of day one was more relaxed, as the irrepressible Steve Broadhead from Broadband Testing leaped onstage and asserted that he hated security - and demanded to know what the industry was going to do about it. The panellists were charged making security too expensive, inconvenient and hard to manage. Naturally given the nature of the topic, there was no hard and fast answer to Broadhead's heartfelt plea.

The event then broke into press/vendor meetings, punctuated by a dinner at the Casa Valha. The next day, with some bleary eyes in evidence, NetEvents Portugal resumed its final plenary session after more press/vendor meetings and lunch. Debate IV's topic asked whether 100 Gigabit Ethernet is ready to roll. Surprisingly, the panel's conclusion was that it is -- more or less. While we won't see mass deployments for a few years, the technology's standardisation is well under way -- it just needs to fall in price from its current astronomical levels.

The final debate before the assembly broke up to head for the airport discussed how data centres could be made greener. While everyone agreed that there more energy could and can be saved, there was also acknowledgement that this had to be balanced with the need for the data centre to be over-sized to allow for expansion, as new technology and tools helps to resolve the problem.

The next events will be held in October in China for the Asian press, while Barcelona, not coincidentally the venue for Mobile World Congress, will see the next European event in February 2009. See here for details.

Is VMware’s star on the wane?

It’s not long ago that VMware was at the pinnacle of the virtualisation industry. And of course it still is, by any measurable criterion. It has the biggest market share by a long way, has the widest user base and the biggest roster of partners -- its famed ecosystem. But as the company this week cranks its VMworld show into action once more, recent events suggest that VMware’s star may have reached its zenith.

When the company started nine years ago, it set out to develop and popularise virtualisation technology, which then looked like a nice-to-have, in that it can cut hardware maintenance costs, allowing the number of admins per server -- a common metric for assessing the productivity of IT services departments -- to plummet.

Virtualisation technology carries hidden costs too. While you have fewer hardware servers to manage, the number of software servers is likely to end up even higher than their original hardware counterparts. Each of these needs to be managed to ensure that they’re patched, the security software is kept up to date and so on.

What’s more when you consolidate, say, one hundred servers into ten servers, the throughput required remains the same as before. This can entail costly upgrades of ancillary equipment such as data centre backbones and other elements of the network. additionally, those ten servers will need to be highly specified for internal throughput, so they won’t be ten times cheaper to buy.

What you then need is virtual machine management software -- stuff that can orchestrate all this, and ensure that backups are kept so that, would hardware or software failure occur, you have a duplicate or mirror somewhere in the mix to get that service back up and running. Sound complex? Yes, it is.

But is all becomes worthwhile once you factor in the big win in today’s eco-aware climate. There are energy, carbon and emission savings to be had by reducing the data centre’s power draw. Plus, many data centres are already at the limits of the mount of power they can consume, so virtualisation is key to meeting the relentless demand for more computing resource.

This makes VMware’s situation at the top of the virtualisation hill a good place to be. Virtualisation technology remains a burgeoning market, with fewer than 30 per cent of the world’s servers virtualised according to one recent set of figures, so there’s a lot to play for.

Yet there are lot of pretenders to VMware’s crown out there. The biggest and most threatening of these is Microsoft. Having just cut the price of its hypervisor product to zero dollars (only from its original US$28), Microsoft is doing to VMware what it did to Netscape in the browser wars of the 1990s. The basic virtualisation kernel will, as VMware freely acknowledges, become free and commoditised, while the value-add -- and the big revenue earner -- will be the management software.

Yet VMware’s successful efforts to build a partner company ecosystem around its hypervisor product could mean that its core differentiator is undermined by the vigorous development efforts of those partners, many of whom are focusing on the thorny issue of managing hundreds of virtual machines in a data centre environment. Niggling away in the background is the knowledge that parent company EMC has effectively fired the two people who started the company: CEO Diane Greene, and CTO and Stanford professor Mendel Rosenblum -- both her husband and the technical driving force behind the technology.

Can VMware stay on top? Will the appointment of an ex-Microsoftie to run the show mean that VMware has an inside track on combating one of the most fearsome and aggressive competitors on the planet? Will EMC regret its meddling with a company that wasn’t broken?

Such questions have to be answered. Yet there’s still a lot of momentum behind VMware, a huge number of very smart (and, for that matter, likeable) people driving the company forward, and a large number of partners and customers who want to see VMware carry on.

But in the long run, will it be enough? Time will tell but it will doubtless be an interesting battle to observe.