Continue reading »]]> The LEO Computers Society celebrated today with a lunch of at the Science Museum. The first regular business application used within the world was used 60 years ago today. This was the LEO computer, built and deployed by J Lyons and Co, to manage the operations of their tea-shops. It completely took over the Bakery Valuation calculations 60 years ago today, having been running test programs since the 5th of September.

LEO was based on the EDSAC design built in Cambridge in 1949. LEO Computers Ltd was formed to manufacture the LEO computer for other companies in 1954. This technology was installed in some interesting locations, such as both the Parts and Payroll departments of the Ford Motor Company, the British Oxygen Company, and the Ministry of Pensions. These all received a LEO/II version of the system, the first system installed in May 1957.

An updated platform the LEO/III was released in 1962, and most of these were deployed within the GPO (General Post Office), with others in HM Customs and Excise, HM Dockyards. Indeed, by this time the government had got to grips with the benefits of computers. The LEO/III was also used by many corporate customers, both within the UK and further abroad, with deployments in Shell in Melbourne, Australia, and a LEO office in Johannesburg, South Africa. The LEO/III was one of the first solid-state computers, and used a ferrite-core memory. Later improved versions the LEO 360 and LEO 326 were manufactured by English Electric after it took over LEO computer in 1963, later being formed into ICL in 1968 with International Tabulators and Calculators.

One of the interesting things about the LEO was the use of sound to provide an indication of the internal state of the system. This lead to some interesting developments, such as the use of this to generate seasonal sounds by using specific programming commands. Perhaps the first use of corporate compute resource for recreational purposes? Listen to this MP3 to find out..

Continue reading »]]> In my previous post “Will the Thailand floods prompt a move to SSDs?“, I posited that the floods in Thailand would increase the price of traditional spinning media so that it would rise above the cost off SSDs (or at least make the move to higher performance storage systems easier to justify.

With the following article appearing in The Register: NEWSFLASH: Chips cheaper than disks, it would appear that this has happened, at least for enterprise class hard disks and SSD devices.

“Recent HDD/SSD distribution indicates an average 20-25 per cent increase in Seagate and HGST’s enterprise-class HDDs (e.g., 15K.5 73/74GB HDDs nearing approximately $3/GB), which compares to OCZ’s 6Gbit/s SAS Talos SSD at $1.98/GB; single-ported Deneva SAS at about $1.55/GB).” – Aaron Rakers

Continue reading »]]> As part of an investigation into several other things, my Internet research dredged up some work by Jim Gettys on Bufferbloat, which was related closely enough to my intended target to get stuck into it a little more.

It has been previously been believed that adding buffers across a network to help mitigate against packet loss is a “good thing™”, especially if we’re going to be dealing with VoIP and other protocols sensitive to packet-loss. Bufferbloat is the phenomenon caused by these buffers when the data source(s) can generate enough data to fill these (often deep) buffers, and thereby destroy the linkage between TCP’s congestion control mechanism and the actual data rate.

This leads to added latency, increased packet loss, lowered network efficiency (all those retransmitted packets take up space that should be used for other data), and lower throughput. These are all situations that we have experienced in the Internet, every day, and is typically shown by the sporadic performance in downloading large files across the Internet.

There is a nice introduction to bufferbloat on their page, which explains the problem, which is basically that each and every buffer within a network is a potential problem, as if data can be sent into the network faster than the buffer can empty, this will have a knock on impact on the performance of the flow (of whatever application or protocol). Whilst this does optimise the sue of bandwidth in the network, (i.e. it keeps it full), it doesn’t optimise the performance of the flows, and this means that the network efficiency drops as it’s carrying many packets that are retransmissions of the original (still buffered) data, because it hasn’t yet arrived.

The solution in the past has been to implement technologies such as random early discard (RED) in order to determine if the buffer is filling, and to selectively discard packets on flows that occupy more of the buffer than others. But it hasn’t always fixed the problem.

Whilst the Bufferbloat project is looking at ways to minimise buffer size (and potentially self tune the buffers within devices), the situation exists in each and every network, because buffers are cheap (both in parts and logic) to implement. But it may take a long time until router and switch (and all the other bits of comms kit) suppliers fix their environments to adjust automatically. The internet might be a challenge for a while, but it’s certainly possible to fix the enterprise environment.

So how do you manage the performance of a network if the feedback mechanisms are broken? I believe that the best mechanism is the management of the performance of each flow against the available bandwidth at the edge of the network. How can we implement one of these?

We need:

• A system with intelligence at the edge
• A way of understanding the available bandwidth in use at each point
• A mechanism to ensure that packet loss is managed to protocols that are less ‘important’
• A mechanism to determine if there is bufferbloat in the network core

It sounds like just the sort of things that the Ipanema Technologies‘ Autonomic Networking System delivers.

Continue reading »]]> I can’t quite believe that the android platform is already four years old. (well I can, since the recent Android 4.0 (Ice-Cream Sandwich) announcement.)

Google were expected to make an announcement about the pending gPhone to take on the iPhone, but instead announced the Open handset alliance, and the Android operating system as a core part of the platform.

Since then we have seen various tasty Android platforms, such as Eclair (1.6), Froyo (2.2), Honeycomb (3.0 for tablets), and others.

The hardware platforms have included the Google Nexus, Motorola Droid, HTC Incredible, and the more recent HTC Desire HD, Samsung Galaxy II series, and Google Nexus S.

My opinion is that the Google platform has some challenges to face. Handset manufacturers are going to be challenged to deliver differentiation between their handsets, as the struggle to build, debug and maintain a UI layer over the top is going to be a challenge; Google’s recent acquisition of the Motorola mobile phone is going to potentially cause splits within the Open Handset Alliance; and the biggest issue is the Apple vs Samsung IP battle.

The next few years are going to be fun for Android!

Continue reading »]]> With ASUS announcing that it will run out of hard disks at the end of the next month [via theregister.co.uk/], since new ones are not being produced in enough numbers to meet demand. In this scenario, prices are rising sharply.

An example of the price increase of hard disks is this WD 2TB disk on e-buyer at £211 pounds currently. (Anecdotally, this drive was previously available two weeks ago at just over £50, a 400% increase!). I brought a pair of these Samsung HD204UI disks last month from Scan for £46.99 a piece, these are now £125 each, a 266% increase.

Given that the increase in hard disk prices has been quite sudden, and dramatic (and may yet continue), perhaps as we see the cost differential between solid-state and hard-disk is reduced the value of the performance difference between the two systems is increased. This makes it easier to see the performance benefits as a substantial improvement, rather than a significant cost.

Perhaps this could be the point at which the threshold is reached for the transition from the old hard-disk media to the newer, more rapid SSD platform start to occur, especially for single-disk platforms such as laptops, and netbooks, and the ultrabook (where an SSD is a necessity to meet Intel’s ultrabook requirements).

For those that need to store more information, it will still be cheaper to buy large capacity hard disks, but for the medium-term, in think that the low cost per gigabyte seen recently will be a thing of the past.

There will still be requirements for large slabs of storage especially for those cloud hosted solutions. Perhaps we’ll see the large swathes of “spare” storage in Google’s gmail become more reduced until production capacity can be returned.

But Thailand isn’t the only place where hard disks are produced, so perhaps Samsung will take advantage of the situation that Seagate and Western Digital face.

Continue reading »]]> The recent acquisitions by Riverbed of Aptimize and Zeus (to provide A to Z coverage?) have now been integrated into the Riverbed family. See “Riverbed goes shopping for godly web performance” for details of the acquisition announcement.

So now as well as the Steelhead WAN optimization appliances, the Whitewater cloud storage optimization appliances, there is now the Stingray product range to add to the mix.

This comprises the Stingray Traffic Manager that frontends web services, and off-loads SSL, caches, and balances traffic. This can be deployed both in the cloud, or in the corporate environment. There’s also the Stingray Application Firewall, which is able to provide web-application firewall capabilities.

Both of these have come from the previous Zeus stable. The Zeus webserver, which has become renowned for it’s ability to perform under load and it’s scalability will no longer be marketed.

The Stingray Aptimzer is from the Aptimize acquisition (the clue might be in the name). This is using the technology to look at the various transfers and to use cached and optimized elements for the transfer. This allows for example (similar to the Redline then Juniper DX devices), the ability to take an image file and optimize it for the receiving device.

Continue reading »]]> Computing is losing a generation of legends. Following the recent deaths of Steve Jobs and Dennis Ritchie, today I learnt of the death of John McCarthy.

From Facebook, Philippe Khan (who I had the pleasure of working for a time) wrote:

RIP John Mc Carthy. I consider John’s original work as defining Artificial Intelligence. Practically we’ve all used and use LISP or derivatives such as Scheme. When John did that core work was before my time, yet I still marvel at the clarity of thought and how modern it continues to be today. Thank you John!

http://www.bbc.co.uk/news/technology-15444222

John McCarthy was heavily involved in the start of the AI (Artificial Intelligence) movement, and use of mathematical logic to deliver it. As one of the proponents of the 1955 Dartmouth Conference, he stated that the problem with encoding intelligence was not the capacity of the computers, but the ability to encode programs that take advantage of the capacity that they have. In addition that conference also discussed Neural Linguistic Programming and Neural Networks.

The LISP (LISt Processing) language was a natural off-shoot of this development, using evaluated parenthesized functions and arguments. LISP is inherently recursive, a problem that caused memory fragmentation, so John McCarthy also gave the world “garbage collection” routines to recycle, amalgamate and reuse blocks of memory.

In 1961, he discussed the future of time-sharing computing leading to a utility-based model for consumption of computing resources, a model that in the last few years has driven the grid and cloud computing environments.

He posted on his own web-page about global sustainability, and backs these with simple numbers gathered from quoted sources, which leads to one of his quotes, which I quite like.

Many questions can be settled by recourse to available statistics and arithmetic… The converse is that failing to look up statistics and do the arithmetic is a recipe for ignorance. source

In 2001, his short story The Robot and The Baby shows how systems which are programmed with goals may give results that are consistent with those goals, but not necessarily what you would first expect.

Again, computing has lost a visionary mind, and advocate of it’s use for the common good.

So long, John McCarthy.

Continue reading »]]> I’ve not blogged about the events in Wootton Bassett over the past few years, in part because I’ve taken a stand (similar to many in our town), that what we do is a personal response to the fallen members of our armed forces as they are repatriated through RAF Lyneham.

Today, the town has been granted Letters Patent from the Queen, and officially becomes Royal Wootton Bassett, in recognition that of the many that gather on the streets to offer our respects to those who have offered the ultimate sacrifice for their country. This really came about because of an accident of geography, and the close links that the town has had for the base at RAF Lyneham over the years. Indeed, a lot of ex-service personnel have retired to the town.

The process started when an RAF Hercules C.130 aircraft of 47 Squadron, based at Lyneham, crashed in January 2005 on active service over Iraq. Since many of the people who work at Lyneham live in the local area, nearly everyone knew someone who was directly or indirectly affected by the crash, and the town gathered to support their friends and families. This occurred before we moved to the town.

Normally RAF Brize Norton in Oxfordshire is the main location for passenger movements, but when that was closed temporarily in the September 2005, to allow upgrades to the runway. During this time most operations were moved to RAF Lyneham, along with the repatriation of the bodies of the fallen from Iraq and Afghanistan.

The first repatriations through the town were low-key affairs, with no formal organisation, just an awareness by those older members of the local Royal British Legion of hearses passing through the town, and the recognition of the service given by the servicemen within.

This grew to a more formal recognition of the repatriations, with an informal network of phone calls and discreet displays in certain of the shops on the High Street of the times of the next repatriation. Those that were on the High Street at the time stopped what they were doing, and stood in silence as the cortèges stopped at the War Memorial.

The repatriations through Wootton Bassett finished in June 2011, as the last flying squadron left RAF Lyneham and moved to RAF Brize Norton. Repatriations moved there as well, with the first repatriation occurring on the 8th of September 2011, through the town of Carterton in Oxfordshire.

The solemn respect that has been offered without any reference to the politics one way or the other for the deployment of our troops abroad. The quiet dignity in which every one in which those in the town who gather on the High Street as the cortèges pass has now become world-wide news.

Today we’ve been recognised by the Queen (and country) for this task, which by accident of geography, we’ve been able to carry out for the 355 servicemen and women who’ve passed through our town. I’d hope that others would have done the same, and will continue to do the same.

In the light of that, our town coat of arms has been changed, and we’ve lost a gold diamond from the below the chevron, and have instead, gained a royal lion, there instead. A fitting mark for the loss of all those diamonds in our armed forces who are prepared to give the ultimate sacrifice in service of their country.

Continue reading »]]> It must be getting to that point in computing’s history, as it appears that the pioneers are passing from the world.

Sometimes it’s all to easy to forget that computing didn’t really start until the early 60′s. Before this point, we had a lot of interesting hardware platforms (such as Babbage’s difference engine), but the science that supported them in terms of maths, logic and electronics were very much in their infancy.

During the 1960′s advancements in logic (the use of a sequence of Boolean circuits to convert input data to output information) were made, both at the theoretical level, and then through the practical level with transistorized logic gates becoming available. These lead to some specific dedicated function “computers”.

But the development of flexible platforms that could be “programmed” to provide specific tasks required not-only the machine-code interface to allow the system to operate at its most efficient, but also higher-level programming languages to allow the task to be broken down into meaning routines, and for these to be coded at a higher level that the machine code for programmer efficiency. This had to wait until the late 1960′s, when a lot of people were looking at the rapidly advancing underlying hardware, and programmers had challenges in making efficient use of the new capabilities of the hardware in a new and novel machine-code for each platform.

Dennis Ritchie was a key exponent of the use of higher level languages to achieve this, and he worked on many over his life time, in particular B and later in conjunction with Brian Kernighan, the C language, developed during 1969 and 1973 at the Bell Labs as part of the Unix platform (indeed, C is an integral part of the Unix operating system, with most of the system itself encoded in C after 1973, having initially been encoded in Ken Thompson’s B language). The most definitive form, the so-called K&R C (published in 1978 as part of the “The C Programming Language”), became the informal specification of the C language, and a definition of who to do it “right”, and have it work on almost any C compiler.

C is not really a high level language, such as Pascal, or COBOL, developed at about the same time. These languages had a high level of abstraction from the underlying hardware. C however, evolved to support the need to get the best from the underlying system platform, and so evolved elements to gain the most efficiency from the limited hardware resources available. (Witness C’s ability to manage bit-packed fields without significant overhead, which is great to manage device registers)

The flexibility of C and subsequently C++ has allowed it to become the standard for most systems languages, and many applications are still built on top of the C infrastructure. Indeed, this is just one of the reasons that he achieved acclaim outside of the world of computing, being granted in 1999, the National Medal of Technology (along with Ken Thompson) for the development of C and Unix, one of many given during his life for this work.

Dennis Ritchie followed the archetypal view of the early 70′s computer scientist, with beard, long hair, and an inventive and agile mind. It’s that last bit that is going to be missed in the future.

So long, Dennis Ritchie.

Continue reading »]]> Just over a month ago, I asked “Has Jobs really left the core of Apple?“. Today, sadly, it is certain that he has.

In the news, pictures show the Apple (APPL) offices at 1 Infinite Loop with the flags at half mast
, and their web-site currently shows a simple message stating

Apple has lost a visionary and creative genius, and the world has lost an amazing human being. Those of us who were fortunate enough to know and work with Steve have lost a dear friend and an inspiring mentor. Steve leaves behind a company only he could have built, and his spirit will forever be the foundation of Apple. – www.apple.com/stevejobs

According to the Wall Street Journals obituary for Steven Paul Jobs, 1955-2011, that he passed away, peacefully with his family, after his fight with pancreatic cancer. The BBC also have an obituary for Steve Jobs, along with UPI (Steve Jobs dies at 56) and many others, including Wired (Steve Jobs passes away, aged 56, on October 5th 2011)

This announcement unfortunately really overshadows for Apple the announcement yesterday of the A5 powered iPhone4S, the hardware platform for the iOS5 operating system which is the real iPhone (and iPad), and provides the infrastructure for the there’s an app for that infrastructure, turning iTunes into something more than just an online music store.

Whilst I’ve never really been an Apple fanboy, slavishly following the herd, and getting the latest Apple hardware, I do recognise that Steve Jobs is probably the most influential and visionary of the initial generation of personal computer companies.

It initially started out with the Apple I, the first computer that Steve Wozniak
and Jobs worked on at the Homebrew Computer Club. This was later followed by the Apple II which was a pre-packaged platform, for those that wanted a computer capable of flexibility and graphics, but didn’t want to waste time with a soldering iron. It also supported an open expansion slot capability which allowed it to be extensible.

On 24th January 1984, Apple launched the Apple Macintosh, with the now well-recognised 1984 Ridley-Scott directed advert. This was the start of Apple’s drive to be a platform for the creation, editing and delivery of content, the start of the Desktop Publishing revolution (with the LaserWriter being the first PostScript Laser Printer being launched in March 1985).

In 1985, Steve Jobs founded NeXT computer, and delivered a Unix-powered cube that again became a game changer, with the object-orientated NextStep operating system. In another claim to fame, a NeXT computer became the first World-Wide Web server in the world, when Tim Berners-Lee created the HTML and HTTP protocols at CERN. Whilst NeXT was not a commercial success, some of the technology and ideas there later become a part of the Mac OS X operating system when Steve Jobs returned to Apple in 1996, when they brought NeXT, in turn allowing the charismatic Steve Jobs to return to the company he founded, bring back direction and coherency of vision to a company that had clearly lost it’s way at the time.

In 2001, Apple launched the iPod, moving the company into the consumer electronics space, and taking the Sony Walkman concept significantly further, with your entire library of music now being portable. Many iterations and form-factors of this technology are available, from the small shuffle to the larger hard disk based classic.

In 2007, the iPod morphed again, with the addition of a much larger screen for movie watching, and more importantly a wifi connection, as well as a mobile data connection. This was the iPhone, probably the most significant of the smartphone platforms, and one that sets the bar by which others are measured.

In April 2010, the iPhone got squashed and flattened, to become the ground-breaking tablet that is the iPad, and it seems hard to believe that it was only 18 months ago that these slim, light-weight, long-lived devices weren’t around, and that tablet used to mean a clunky laptop-type device that was heavy and didn’t have the battery life to support the things you wanted to do with it, even if you could carry it around with you.

But Steve Jobs wasn’t all about technology, he was very aware that the technology needed content to push it, hence the iTunes store deals with record companies, movie deals when the iPhone arrived, and the in-built connectivity of all Apple computers since the Macintosh for collaboration.

This also drove his purchase of Pixar, to my mind. Pixar would almost certainly not be in the position it is today with out the backing of Steve Jobs. Pixar certainly took advantage of the NeXT computer platform as part of the Renderman software specification for turning collections of individual models in to a stream of film frames.

This symbiotic relationship helped strengthen both companies, but Pixar emerged the stronger, surviving the hardware, proving again that content is more important than the delivery platform. A concept that Steve Jobs knew intimately and shaped lots of his decisions in life.

Steve Jobs, thank you!