Last week we ploughed through the Friday afternoon traffic to Nortel’s Maidenhead offices to take a look at WiMAX in action and meet the people behind the new alliance.

Friday afternoon traffic is much like any other network; bottlenecks caused by roadworks, accidents and plain old poor planning, then a long, empty stretch of motorway when you finally get to it. Think of the motorway as the Internet backbone and the A and B roads as the mess of routers, wireless access points, Ethernet and DSL that gets you online in the first place.

The connection in our office was more like permanent roadworks than an accident; you get used to it being a bit slower than you’d like. We upgraded to DSL Max some months ago and noticed an improvement but it wasn’t till we put in a new router designed for ADSL 2+ connections rather than the ADSL 2 hardware we’d had hanging around that the traffic cleared.

Loading Web pages seemed a bit faster, but it was a DVD image download from MSDN that showed the real difference - from the old 400Kb/second to over 750. A speed test puts the possible throughput anywhere from 4,414Kbps during the day to 6,350Kbps in the early evening.  We’d upgraded the network bandwidth, but not plugged it into something that could take advantage of it.

The WiMAX demo was impressive. Watching Ellen deGeneres pet tiger cubs on live TV courtesy of a Slingbox and a Fujistu ultraportable was like watching Freeview rather than YouTube; the video was smooth and high quality, even on a difficult image like fur. Next we tapped into the video feed from the Toshiba HD surveillance camera swivelling around on the ceiling. This wasn’t transmitting in HD - 720p would take 20Mbps rather than the 5Mbps that WiMAX chipsets can currently handle, but we could still zoom in on a poster 20 feet away and read it or see what was on the screen of a mobile phone on the other side of the room.

That works well on a notebook with the screen size and the software to watch full-screen video. You’re not going to get quite the same experience on a mobile phone. One of the key things with WiMAX will not just be the speed of the network and the cost of the bandwidth, but what you can connect up with and what you can do on the device.

Will you be happy with a smartphone and a handful of apps, or will you want a real PC that fits in your pocket?
-Mary

There’s an Onion article that chortles about a supposed internet outage, and its effects on social network addicts. It’s a spoof that nearly came true earlier in the week, when one of San Francisco’s largest data centres (conveniently located to the SoMa dotcom hotspot) went down.

It’s an interesting study in risk management. Most of the services that went out weren’t running globaly redundant data centres - so with all their machines in one site, a single fault took out an entire service. A DNS failover would have allowed a business with a redundant site to continue operating with minimal downtime. Instead we were left with photographs of queues of sysadmins stuck outside a data centre, waiting for its security systems to come back online.

So why did things fail so catastrophically? A big data centre like that is going to have some form of short term power storage to cover the period while a bunch of large generators start up. At least we can be assured that that was the case here. The building actually had some of the latest interim power systems - in the shape of several large flywheels. Spun up while there’s power, flywheels can store a lot more energy than the equivalent volume of batteries. In most circumstances more than enough to keep a whole centre online while the generators autostart.

Unfortunately things didn’t quite work out. The power to the data centre had been going in and out for most of the morning, in a series of glitches that weren’t quite long enough to trigger the generators. In between times the flywheels weren’t able to store the amount of energy they’d given up. When the power finally gave up the ghost, there wasn’t enough energy in the flywheels to power the data centre - or to trigger the autostart on the generator.

The result was inevitable. Servers crashed (or at least used their own UPSes to shut down gracefully). Data was lost, and users were left staring at empty browsers. The refresh button wasn’t working…

So what lessons can we learn from this? One thing is clear. The old rulesets for generator autostarts and transitional power systems need to be rethought. If flywheels or batteries are running low, then charging should be a priority. If power glitches are still occuring, then generators should be running ready to go to full power at a moment’s notice. It’s also probably a good idea to mix transitional power systems. Flywheels are great, but keeping a room full of batteries as well is probably a good thing. When you business is other people’s businesses, it’s a good idea to have both belt and braces ready.

Customers also need to be ready for outages. Global redundancy may be expensive, but it’s critical for many online-only businesses. It’s also clear that despite the assurances of your hosting providers you’ll need to implement your own UPS - even if it’s only to shut down servers gracefully. It’s also a good idea to have a restart plan - making sure servers come up in the right order to get your business back on an even keel as quickly as possible.

Outages like this can teach us all a lesson. Risk management isn’t something that can be ignored, even when you think you’re protected. It pays to expect the unexpected - and when it happens, to respond in a manner that complements your existing automated systems.

First we saw mobile phone networks sharing base stations, instead of building them side by side. Now T-Mobile is using BT for it’s backhaul - the connection from the T-Mobile cellular network into the Internet - instead of building a network of its own duplicating what BT already does.

T-Mobile gets a virtual private network on the BT lines - just as Virgin gets a virtual mobile network on the T-Mobile network. Originally Apple planned to do have its own virtual mobile operator for the iPhone, but you have to employ a lot of customer service representatives, have a lot of phone lines and be ready to take the blame as well as the praise for the performance of someone else’s mobile network. When AT&T accepted a variant of the deal that Verizon turned down, which includes giving Apple a percentage of the revenue for all the calls made on iPhones, it worked out to cost Apple as much and make Apple as much money as running an MVNO on the AT&T network - but without Apple needing to do as much of the work.

Like using Amazon’s storage or Salesforce.com’s servers or Microsoft’s MapPoint service, you’re paying for the benefit of an infrastructure that your business doesn’t have to build and maintain. This is the network equivalent of Web 2.0.

The deal is also good news for BT, because it proves that there’s money in building a high-capacity 21st century network. Initially the service will run over existing leased lines but as BT builds out the network T-Mobile will move over to Ethernet. And that means converged mobile/fixed/VOIP telephony and complex network services all get easier - and possible cheaper. That is one big difference from what Apple is doing with the iPhone - which is limited to simple data connectivity on the EDGE network.
-Mary

This blog entry is going be written using 160 characters.  That’s not a lot, but we’ve been sending messages this long for 15 years: today is SMS’s anniversary. At least I’ve not used any TXT SPK.

We should see the beta of SP1 for Vista some time this week - maybe even today. My crystal ball also tells me we’ll see complaints about what is and isn’t in there, criticism about how soon or how late it is and a few brickbats for the idea of having a service pack already. Given we’ve heard so much about how companies like Intel still don’t want to move to a new OS until SP1 comes out, Microsoft can’t win. People are still complaining about Vista, from the price in the UK (rip-off Britain is a reality) to the perception that it’s XP in a new frock (er, no; from the kernel up, it’s new, different and mainly better and at the very least SP1 will make group policy easier to work with).

If you want to criticise Vista, criticise it for taking so long to deliver something that’s good when it should have been excellent; criticise it for refusing the hardest fence (the object file store - the Hamlet’s father’s ghost of Windows); criticise it for leaving out key security technology (BitLocker) for all but Software Assurance customers and people shelling out for Ultimate.

The argument that came up a while back about whether Vista had too many ways to shut down shifted into an argument about which features should keep getting redesigned during an SO build. For me it’s a question of which teams have enough/insufficient/undue influence on the components of Windows; how do you train team leads halfway down the tree to recognise systemic problems and how do you make the upper branches smart enough to act on the problems? And there’s my current favourite: why is such a big place as Microsoft with so much money filled with people saying ‘yes, but my team is too small and doesn’t have the resources to actually finish our product properly the way we’d like and you the customer are asking for’?

There’s a huge disconnnect inside Microsoft between the incredibly smart people doing neat things who haven’t yet been snapped up by the engineer predator that is Google and the bland range of what actually ships. Microsoft is making software for the mass market but the mass market is moving closer to the edge every day. Enterprises are rolling out Vista at the same pace they roll out every new version of the OS - but small business is a bigger market than enterprise. It’s a harder market - if you want to sell 1,000 licences you have to make 1,000 deals not one or two. And like the home users who turn into office workers when they get on the tube, small businesses are willing to give things a try. And one of those things they try -that might be the next big thing for enterprise and conservative user alike.

So there’s a good reason for shipping far more of the cool stuff. The problem is that when Microsoft ships stuff, the expectation is that it’s finished bits; polished and it not bug free then supported. Live is a confusing name. Live labs is even more confusing. And software plus service is an excellent notion, but moving forward means addressing the software side as well as the service side.

So if they need a new label for releasing nifty things, why not Microsoft 2.0?

Microsoft 2.0 could take things like Live Labs and add in the hundreds of interesting research projects like Aura and SNARF and the podcast editing project that lets you edit sound by selecting, moving or deleting words in the linked transcript (with the snappy name of “Improved Podcast Authoring with Speech Recognition”). I’ve met people at Microsoft who say that’s what their job is, but it’s a struggle to get past the culture.

So maybe Microsoft should set up an incubator for projects that would be a whole Web 2.0 IPO anywhere else; give the developers a budget and workspace on campus (so they’re still getting to talk to the other smart people working on other software) and 18 months and a guaranteed job to come back to and let them run with an idea. At the end of 18 months either the team gives up and moves on or Microsoft has a successful new product to move into the development and support cycle.

It’s the way IBM created the PC and all the subsequent successful PC divisions, although IBM also gave each child project permission to throttle parent company and use it as compost. It’s the way new divisions like X Box and Windows Mobile have worked, but with anointed leaders like J Allard and internal bickering; so far Microsoft hasn’t managed to let anything as dear to its heart as software get that far out of the core teams. There are many neat things to be done and many neat things that Microsoft is doing. It can either find a way to join those two sets of dots or turn into the next Xerox Parc and wait for the next Bill Gates and Steve Jobs to show up.

-Mary

You can tell I’m a technology journalist. It’s not the perpetually dishevelled look, the scrawled notebook, the geeky t-shirts, or the laptop covered in stickers, though I’m sure they’re clues.

No, it’s the regularly changing collection of backpacks that really gives things away. Nearly every big event you go to, there’s a new bag to take away. Some people might be satisfied with that, but I’m always on the lookout for the bag that I can use every day. Most I try for a while, but then go back to one of the old standbys. In my case those are a battered old Intel IDF back packand a Microsoft PDC shoulder bag.

I’m sure our local Oxfam dreads our arrival with a car load of conference bags from all over the world. However, if we didn’t pass them on to charity our office would quickly fill with unwanted bags. So how do we decide what to keep, especially in these days of ever decreasing baggage allowances. I tend to classify them as good (worth trying for a while), bad (straight out the door - sometimes into the fabric recycling bins), and indifferent (which wait for the next charity shop run).

We’re recently back from a long US trip, following the conference circuit for a month, and picking up bags at nearly every event we visited. Most of them had to be left in the US, but a few came home with us.Now we often review hardware and software - so why not bags? Some conferences just used plastic bags, but others went the whole hog.

Microsoft MEDC had a decent back pack, with plenty of space. A non-descript black, it also has a magnetically sealed phone section that works well for MP3 players as well as phones. There’s some padding for a laptop, but it’s not one I’d drop from a height. All in all,a decent rucksack with some staying power. The small logo also helps. It’s one that made it all the way home.

Microsoft WinHEC sadly had an indifferent bag. There was plenty of space, and the ability to switch between back pack and shoulder bag modes. However the zips were poor, and it was also pretty ugly. A lack of padding makes it not really suitable for carrying laptops. Still, it worked for carrying food on a road trip. However, the bag itself was quickly retired.

Microsoft TechEd was much the same, down to using a variant of the WinHEC bag, but made from a much nicer (and stronger) material. However, the overall design still isn’t particularly suitable for a working backpack. However it did sterling work as emergency hold baggage for the flight home.

Where 2.0, O’Reilly’s mapping conference, had quite a small bag, a comfortable and well-padded single-strap bag that works quite well as a day pack, with enough room for a book and a camera. It’s too small for anything but a UMPC, so fails as a laptop bag, but will replace an existing bag as my weekend day pack.

HP’s recent events in Las Vegas also added to the bag piles, with Software Universe coming up with an attractively coloured day pack that worked well as basic airline carry-on. However, it’s not a keeper - especially as the parallel Technology Forum even came up with one the the best conference bags I’ve seen in years.

The best bags are generally bags from a existing bag brand, not something sourced from a conference bag supplier. HP choose to use High Sierra brand bags, with just a small logo added to the front of the pocket.There’s plenty of space, and compression straps to make sure that when empty it doesn’t expand to fill a tube carriage. The laptop section has two zips, one giving access to the whole section, and one for getting a laptop out for scanning through security. It’s also well padded, and strengthened. There are several more sections, and two water bottle carriers. It’s comfortable and can hold a laptop, plus all my usual gubbins without removing the compression straps. Disconnect them, and you can add a trip to a supermarket or a hefty camera. All in all, this is one of the best laptop bags I’ve used, and it’s quickly become my regular bag.

So what’s the best bag you’ve come across?

– Simon

I manage my Windows Mobile alarms and alerts pretty aggressively because nothing says ‘you’re out of your time zone’ quite as annoyingly as an alarm for something pointless waking you at 4.30am. I switch the phone automatically into flight mode overnight to avoid late night calls and I have it switch to silent mode when the radio turns on to get email early in the morning, with the sound only coming back a few minutes before the alarm goes off. On our last trip, the one night I left the phone because I was expecting a text overnight, it started howling for my attention at an unearthly hour, Las Vegas time.

The screen was urging me to “Backup Business Contact Manager!” Now; right now; right this minute! But I didn’t set a reminder for that, and if I had I wouldn’t have set an alarm for it. BCM did it all on its own on the new laptop I just set up. And because it’s part of Outlook it pushed the reminder over to my phone when I plugged it in, all without asking.

Backup is good. Backup reminders aren’t a bad thing of themselves. But when you add them to my calendar without asking and without letting me say where, when and how I want to be reminded, you’re shifting the responsibility for not having designed backup into your product properly in the first place. And you’re ignoring the environment your product lives in.

Like The Matrix, too many technologies make up for their own shortcomings by enlisting humans to do the things the computer - or in many cases the programmer - finds difficult. Amazon’s Mechanical Turk makes a business out of it. If I spent my day responding to the pop-ups, alerts, alarms and other suggestions of a fresh Windows install, I’d never get any of my own work done. I want software that’s efficient at looking after itself and that makes my life easier - not the other way round.

-Mary

A while back we remarked on our preference for power supplies that had mini-USB connectors. There are plenty of reasons for looking for a standard wall wart connector. For one thing there’s no crawling under a desk looking for just the right cable, and for another, well, it’s just less to pack and carry in these days of decreasing luggage allowances.

As Sun’s Simon Phipps says, “I now have a growing collection of useful things that work with this USB power standard, all from different places and all interchangeable. So I have a car-socket-to-USB plug, wall-warts for US & UK that deliver USB, and each new gadget that comes with a USB/mini-USB lead makes it easier to leave a cable ready everywhere. And since most power comes through my computer, there are fewer wall-warts left plugged into power sockets acting as electricity vampires.”

Reducing complexity makes peoples lives easier, and standards help control complexity. In these green days it’s sensible to standardise on one power supply so engineers can design the most efficient PSUs possible, rather than having to come up with something new every time a new gadget is released on an ever gadget hungry world. Not only that, but in an emergency we can charge devices from laptop or desktop PCs. And of course if everyone had the same power supply connectors it would make it a lot easier to borrow a charger in the office…

Some gadget companies are worse than others. Apple’s initial iPods sensibly used FireWire for power and data, while the first shuffles plugged straight into a USB port. Now the Shuffle uses a modified headphone socket for power, and the iPod connector gets more complicated with each generation. It’s USB at one end, so why not USB at both?

Nokia is another company that makes things hard for end users. Its power supplies are getting smaller - but so are the plugs that feed power into the devices. And what’s that right next to the power socket on the latest N95? Yup, you’ve guessed - a USB port used for nothing but data. I won’t go into the awful power supply connectors from Motorola and Palm, as I fear for my sanity if I think about them too much.

Well, it looks like the tide may be turning in the favour of mini-USB chargers. China’s Ministry of Information has just made it illegal to use anything else with new devices. The law is initially targeted at mobile phones (upsetting both Motorola and Nokia). However, there are apparently already 15 approved PSUs from 12 different companies, so it shouldn’t be too difficult for phone manufacturers to source power supplies.Switching to one standard PSU format should also save money, and reduce the number of chargers just thrown away after a device upgrade.

There’s really no reason now to stick with proprietary connectors. The largest emerging market has made the decision for anyone designing hardware - so now we just have to wait for the hardware to ship.

– Simon

We’re living in the middle of the digital dark ages. It’s easy to see why, too. Over the weekend we were clearing out the spare room, and a huge proportion of the detritus that went into the black sacks was unmarked CD-ROMs. Some were old betas, some were old backups. There were even a few tapes - including a DAT backup from a long gone Windows 95 desktop PC. Even if I still had the DAT drive (and that went to CiX’s auction conference a few years back) I wouldn’t have been able to do much with the data.

My data isn’t that important in the scheme of things. Most of my files have migrated from machine to machine over the years, and I’ve updated file formats with successive releases of my usual applications. With the latest indexing tools I can find files I created a decade ago in seconds.

But what if you’ve got millions of documents to deal with, documents that need to be kept in the format they were created in, and need to be available to anyone who wants to see them - when they want to see them. That’s the National Archives’ problem. More and more government information is electronic - and more and more information is being created every year. It’s not just the volume of information, it’s also the number of tools and technologies that are being used. The number of obsolete formats is going to keep on increasing - especially the binary formats from different word processors and spreadsheets.

That’s why a wet July morning saw a gaggle of journalists crossing the landscaped lakes of the National Archives’ building in Kew. It’s one of the most important buildings in the country, and one of the least well known. Why were we all going there? Microsoft was going to announce a new partnership with the National Archive, intended to help preserve information held in its many different file formats.

The answer to the problem isn’t converting it all to new “open” file formats. That option wasn’t on the table, as the National Archive has a statutory duty to preserve information in its initial formats - so this wasn’t another pitch for OOXML (or even for ODF). Instead Microsoft is working with the National Archive to provide a library of virtual machines running in Virtual PC. These are initially being populated with different versions of Windows and Microsoft’s Office suite. Instead of opening up a file in 2007 Office on Vista, you’ll be able to choose a VM image from a library, and then open the file inside the virtualised PC. Initially this is just for Microsoft’s own applications, but there’s no reason why this approach couldn’t be extended to any file format stored in the National Archive - from CAD packages and GIS systems, to legal documents written using Word Perfect, and to Lotus 123 spreadsheets.

Virtualisation is an important tool for many businesses - and records management use for it that most companies haven’t thought about. However, as regulations become stricter in their data retention requirements, it’s something that we’re all going to have to think about. Can you still open that pension plan spreadsheet from 1993? It might not be something you need to look at today, but when the Inland Revenue come calling, you’re going to need to be able to show them that not only do you still have the file, you’re actually able to use its contents.

Clearing out our old software we threw away a whole pile of different applications and databases - a few versions of FileMaker Pro, a handful of Offices, a brace of LotusWorks and a small pile of Word Perfect Office. That’s just our collection of old software - how many are sat on the shelves in and around your machine rooms or IT departments? It’s time to start building and maintaining the VM images we’ll need, as when the authorities come knocking we won’t have time.

Working with legacy file formats is a problem that’s not going to go away - and it’s going to get worse before it gets better. How can we work with the ever-changing contents of government wikis, or internal IMs, or the Parliamentary Whips’ pager messages that drive MPs to the ballot?

Vernor Vinge’s award-winning SF novel “A Deepness In The Sky” introduced the concept of the software archaeologist, delving through layer upon layer of virtualisation to find an essential thousand year old file. The National Archive is having to become those archaeologists, with digs into 25 years of online government.