As you may be able to see in the highest-resolution version of the snapshot above (click to enlarge), it’s not every day one comes across a physically distorted DIMM.

This is one of a set of eight 4GB sticks, originally intended to boost the performance of a Hyper-V host machine at Ratcliffe & Brown Wines & Spirits, the subject of a forthcoming PC Pro Business Clinic. The server upgrade wasn’t part of the subject, but it pretty quickly turned into a source of aggravation – this bendy SIMM is not immediately apparent until it’s placed on a flat surface, and I tend to land DIMMs on a lump of textile, like a mouse mat or a rucksack; anything but a conductive perfectly flat plane like a rack-mounted server lid.

Surprisingly, it sat in the DIMM slot perfectly well. Unsurprisingly, the server (a Dell PowerEdge 2970) spat the dummy the minute power was restored, quite accurately complaining about “unusable memory” in the scrolling front-panel display.

I don’t see many DIMMs that wobble to and fro when put down, or double as a leaf-spring the other side up

There didn’t seem to be a “too much torque on DIMM slot 8″ message, even though this was the source of the problem. I’d felt something a bit peculiar with the memory in my hand, but decided it was worth a try: in this particular case, the server was going from 16GB up to 32GB, and if one stick was faulty then we would have to take out the other half of the matching pair, reducing the available RAM to a comparatively meagre 24GB.

Once back out of the machine and laid on a flat plate, the bend became detectable: I don’t see many DIMMs that wobble to and fro when put down, or double as a leaf-spring the other side up: what amazes me is how it got through the memory vendor’s QA processes. Perhaps their tester has jaws that clamp the edge connectors firmly enough that the sundry distorted open-circuits on the stick are forced shut. But it’s visible even when the DIMM has gone back in the packaging, and doubly so when stacked in a pile with other, perfectly normal DIMMs.

Which leaves two questions: how on Earth do you put a permanent bend in a thing as strong as a plank of fibreglass? And, secondly, how did it manage to leave the factory?

Lots of services have had their public IP addresses updated; I am getting calls from clients whose internal systems don’t genuinely use a domain name to get to a service. It’s not uncommon for all manner of software products (including router firmware) to let you type in www.pcpro.co.uk, and then look it up at that moment and convert it to 212.100.242.151 – which is what they then store for future connection attempts.

When we decide to change that underlying server address – which isn’t a bad thing to do, it’s a supported and allegedly seamless choice for a connectivity person to make – these various bits of software and hardware that use “one-shot lookup” simply fail to re-connect the PCs behind them.

According to a few webmasters, this effect includes some of the mass-market BT Broadband routers – you may well see a little note on some sites telling users to reboot their routers in order to bask in the full feature set of the site, for exactly this reason.

Of course, if you have gone away on holiday as all this has started, and you’re expecting to use remote access, or for your servers to carry on being available when they are behind a device which stores numeric IPs after you’ve typed in a name-style address, then you are out of luck. And you wouldn’t necessarily be able to test for this effect when you chose either your ISP or your router and firewall hardware.

On the bright side, a cold restart seems to sort out the problem, in most cases, so you won’t have to give the house-sitter, dog-walker or next-door neighbour the management interface password to your border device, while standing pumping Swiss francs into a callbox at the top of the Spluegenpass (or wherever).

Pull the power lead, count to 20, bang it back in, wait for the little red light to go green.

Comparing the R510 with the old machines is hard, because the simple physical similarity wrongfoots you when you actually absorb the statistics. I gave away 5 LP2000R’s – the virtual machine images of them all would fit, and run, inside the R510 without complaint, and use rather less than half of the current required by just one LP2000R.

Dell have gone for the middle ground with the layout of the R510 – it’s shorter than the normal run of 2U servers at only 24 inches, but manages to wedge 8 drive slots and a pair of Gainestown Xeon 5500s inside that space (Intel are seriously pleased with themselves over Gainestown and the supporting chipset – see Intel’s server video, with chicken).

Usefully, for me at least, the drive slots are sized and structured to take boring old SATA 3.5 inch form factor drives – while one might aspire to HP’s 2.5 inch SAS standard, if cost is a priority it’s nice to be able to use mass market drives in a box that fully understands the upkeep of a RAID array to allow the use of drives that actually do qualify for the ‘I’ in RAID…

I expect to use and abuse the R510 in a couple of forthcoming experiments – some planned, others inadvertent: as soon as I started the box I realised it wanted hardware-specific drivers for the install of Windows 2008, and that the Internet Explorer security configuration had been tweaked slightly to include the Dell site as a trusted destination – but the Dell support automatic model lookup system used web components that IE security configuration didn’t like. Oh, and I have to go and get a heater for the basement – the LP2000R’s used to keep it warm down there, and the R510 doesn’t.

For more details, head to the Dell PowerEdge R510 product page.

It wasn’t the most surprising launch of a server processor ever, but server events aren’t usually renowned for thrills and spills.

Some interesting figures emerged though. They show in just two slides the market reality behind Sun being swallowed up by IBM a few weeks ago, and the sheer dominance of Intel when it comes to processors in everything from enterprise servers to netbooks.

With apologies for the shoddy photography, take a gander at the image below, and click on it for an enlargement if you’re struggling with the fuzziness:

Intel’s Tom Kilroy used this slide to show how the server market in general is growing, which it plainly is, from something like 250,000 units shipped worldwide in 1990, to around 8 million today. But it’s more interesting if you note the key: the blue part of the graph is Intel’s share; the red is competitors’.

The interesting part is that if you squint at the bottom-left of the graph you’ll see it’s all red up until 1992. In other words, back in the late 1980s and early 1990s Intel wasn’t selling any processors at all into servers. Bring it up to 2008 and you’ve got 7 out of 8 million – about 87% – with Intel inside (**BONG!!** bong bong bong!).

That’s a mighty fall for the old purveyors of RISC (as opposed to Intel’s CISC) architecture and what used to be a high-value, high-margin business. It’s always been one of Sun’s main markets.

If you look at another slide – even bearing in mind this is Intel’s own marketing material so was never going to come out in favour of the opposition – the future of non-x86 CPUs is looking decidedly ropey:

A tenth the cost and almost three times the performance of IBM’s Power6? Even taking the required pinch of salt into account, it’s hard to argue with numbers like that.

Intel’s commoditisation of the server market has followed unstoppably from its domination of desktop PCs. The economies of scale in using the same processor architecture and technology in servers, desktops and laptops seems to be pushing x86 unit cost down to a level that nobody else can begin to match. Add to that the compounding effect of the ever-increasing R&D budget that comes of selling so many processors – Power6, for instance, is still on 65nm fabrication when Intel is steaming ahead at 45nm – and you’ve got a terrifying level of domination.

In fact the only place in the market where Intel – or at least x86 architecture – is not dominating is at the very bottom end. With mobile-phone processors and ultra low-power embedded systems, the likes of ARM and Motorola are still hanging on.

In a conference call a late last year, Intel’s Pat Gelsinger stated it would be “decades” before Intel could displace ARM in very low-power devices.

Time will tell whether that was actually a very conservative estimate. I certainly wouldn’t bet on it.

When the weather is hot, and I’m standing in people’s server rooms and they are going nuts with fear and loathing about their precious servers going into meltdown, I like to ask them about Miso soup: and if they get all confused (and then angry and then don’t pay my bill), I ask them how much they think the air inside the typical hot-air balloon actually weighs.

Very few get it right: the answer is, about five tons. Once you get that idea into your head, getting emergency cooling for servers sorted out starts to make a good deal more sense – and those elephant’s-trunk so-called aircon units which harassed managers tend to put in as a reflex action during these periods, start to look more like a way to throw kilowatts into the air for very little benefit, than like a smart way to stop your servers going into meltdown.

Got any good “boy stood on the burning deck” stories from extreme heat or wild weather?