You are here


Computer Storage By Dave Shapton
Published July 2001

Seagate's new 180Gb Barracuda drive is a giant amongst hard disks — for the moment...Seagate's new 180Gb Barracuda drive is a giant amongst hard disks — for the moment...

As technological progress continues to make digital storage devices bigger, RAM cheaper, and processing faster, hitherto outlandish ideas come closer to being realistic possibilities. This month, Dave Shapton thinks about teaching his computer to Name That Tune...

A hard disk landed on my desk a couple of days ago. It had been sent by Seagate for our business to evaluate. Nothing unusual in that: as we supply non‑linear video editing systems, we get through mountains of disk drives. This one looked just like all the others, if a bit bigger and heavier. I was just about to pass it on to our engineering department when I realised that this beast had a capacity of 180Gb.

Big Ideas

The Secure Digital Music Initiative seeks to develop technologies for preventing copying of digital music, and thus safeguarding copyright. However, its four proposed watermarking methods were compromised by a team of Princeton researchers.The Secure Digital Music Initiative seeks to develop technologies for preventing copying of digital music, and thus safeguarding copyright. However, its four proposed watermarking methods were compromised by a team of Princeton researchers.

For those of you who have better things to do with your life than keep track of incremental increases in hard disk capacity, the current generation of disks maxes out at about 75Gb — so this new Seagate is significantly more than twice that. What's more, put six of them on a SCSI chain and you've got over a Terabyte of capacity. Storage is getting cheaper at a faster rate than ever before. Ten years is a long time in the history of computing, and hard disk capacity has grown a thousand–fold in that time: a factor of 10 every three years. I don't know what we're going to do with all this storage but, at the very least, in three years time you won't have so much need to compress your audio files into MP3 format!

Of course, storage is not the only factor driving computer technology forward, although it's an important one. Processing, RAM, and bandwidth are all significant too, and they have got faster, cheaper and bigger (respectively) at least a hundred times over the last ten years. Add all these factors together and you've got a revolution. Multiply them together and you've got a hyper–revolution. That's the only term I can use to describe our state at the moment.

It's difficult to comprehend change when it's as drastic as what we're seeing now, as the Internet pervades our everyday life and two‑year‑olds draw their first pictures on a computer, not on paper. I don't have an exact formula, but I do believe that as the individual elements that make up a computer grow in their own capabilies, the scope of what we can do with the technology grows even faster than these components. Here's what I mean.

As disk capacities grow, we can store more and more stuff. And as processing speeds go up we can do more with the stuff we can store. So having more room to store things is progress in itself, but add to this an ever–growing ability to process the material in the store, and what you actually get is something far greater than if you just 'add' storage and processing together. So let's imagine we've got a massive store of music on our computer. So much, in fact, that's it's getting difficult to find anything. What better way to find your song than singing it to the computer? You'd simply say "find me the song that goes 'dum di dum di dum'" and, in no time, your computer would find the very song for you. To do that, it would have to analyse your feeble attempts at singing the song in question, extract the sequence of notes, and then compare this sequence with an index of melodies it has extracted from the recorded songs, using complex DSP.

This is an extremely processor‑intensive task, but at some point in the very near future, I wouldn't be surprised if such a facility were to turn up as standard, in (say) Windows 2003. (I work with a guy, a digital image analysis specialist, who — in his spare time — develops algorithms for extracting note information from polyphonic audio tracks. He is confident that he is close to being able to do this.)

If that's two or three years ahead, what about 10 years ahead? Storage will probably be a thousand times cheaper than it is today. Processors will be between a hundred and a thousand times faster. Bandwidth will be virtually unlimited and effectively free (although, if BT has anything to do with it, we'll probably still be using 56K modems!) and I can't even begin to predict what we'll be doing with computers then.

Napster Nipped

The storm over Napster rumbles on. There are some new developments, widely reported in the popular press, which suggest that the Peer‑to‑Peer saga is entering its end game.

Under unbearable pressure from the actions of the music industry lawyers, Napster has agreed to 'filter' its content to prevent swapping of copyright works. Using lists of copyright protected songs supplied by record and publishing companies, Napster can now tell whether a request for a song should be processed. In practice, this means that if you search for works by, say, George Michael, you will be told that "no files were found". If you can't find them, you can't download them. If you search for works that haven't been published by a major organisation, the chances are that you will still be able to find them.

On 9th May 2001, ZDNET (a leading computer news source) reported a "precipitous drop" in the number of song trades through the free song‑swap service. It now seems that the loyalty of Napster's customers may be proportional to the number of copyright works it has listed on its servers (remember that Napster doesn't itself store the works — it only points users to the other user's computer where the required file can actually be downloaded).

Napster is, for now, defeated. But it changed the world. And it really did get the attention of the music business.

Since we're talking about copyright, I thought I'd bring up the subject of watermarks. A watermark is a digital 'signal' that is embedded in the audio stream and designed to be immune to processes such as MP3 compression and web streaming. The idea is that if you try to record a watermarked file to a device that recognises watermarks, the device will block the recording. The jury is still out as to whether watermarking actually degrades the sound. In theory it must. In practice — although I disagree with the whole idea of watermarking — it's very unlikely you would be able to distinguish a watermarked work from the original, clean version.

The Secure Digital Music Initiative, an organisation whose aim is to devise methods of maintaining a semblance of copyright protection in the digital age, endorses the use of watermarks as a way of identifying the copyright status of musical works. In an attempt to verify the robustness of their watermark technology, the SDMI decided to demonstrate it in the most public way. They issued a challenge: defeat the watermark and win a prize! It only took a few days for a Princeton Professor, Edward Felton, and his team to crack all four of the watermark methods in the challenge.

As has been widely reported in the computer press and on the Internet, rather than take his prize, Felton announced that, in the interests of research, he would publish details of the techniques used.

Some time later, shortly prior to a conference where he was to reveal his research, he received a letter from the SDMI suggesting that legal action would be taken if he published details of their watermarking techniques. Disturbingly, the SDMI hinted that they would consider invoking the provisions of the Digital Copyright Millennium Act, a —<sub> </sub>shall we say — 'surprising' piece of American legislation that appears to make any action that could lead to violation of copyright an offence.

This is where we can begin to prove that there is something completely wrong and unsustainable in the music industry's fight against the new paradigm wherein computers, as well as gramophones, can play music. You see, the first sign that an argument is invalid is when you can show that if you push it far enough it leads to contradictions.

And that's just what's happened here. I'm no expert on American law, but my understanding is that one of the most fundamental rules of American society (possibly almost as important as the right to carry the means to kill your fellow Americans) is the right to free speech. That's the right which is being violated by the SDMI's proposed invocation of the DCMA. It's also what is preventing me from telling you where you can find all these details on the Internet. Just as well there are search engines...

Beware The Bear: File‑Swapping Dangers

The Napster idea is actually a very simple one. It's one of those things that seems obvious once it's been done. And to prove this, there are new Napster clones springing up every day. I've written before about Napster clones. Some of them don't use a central database, but set up a distributed architecture in which searches are made between end‑users' computers. And it works, if not as well as Napster. (Not surprising, when you consider that the only sense in which the ad‑hoc network is organised is that it uses a specially devised protocol. Gnutella is not a 'thing' in itself.)

But if you are tempted to use services like Gnutella (whose friendliest version is in the form of the Windows‑based BearShare), there are some things you need to know.

First, the Internet is a nasty place. You will find that MP3 files are not the only things you can swap. If you load the 'monitor' window in the BearShare application, you will see that requests for pornographic material run a close second to music requests. If you are easily shocked, don't select the monitor option.

Second, er, the Internet is a nasty place. When you install the current version of BearShare software, you also install software that gets into your computer, monitors your online activity, and sends you "appropriate" advertising. To do this, it probably sends data about you to... well, who knows? And it does this whether or not you are running BearShare. If you find irritating mini‑web pages popping up unexpectedly, you probably have this software (known generically as 'Spyware') on your system. Worst of all, it's difficult to get off your computer because there's no uninstall option.

I now think that the only safe way to experiment with the Internet is to have a separate, dedicated computer. Not only should it be isolated from all other computers you may have, but you must be prepared to format it and re‑install the OS occasionally. And, of course, you should have personal firewall software installed. Since I installed BearShare, my Internet connection has slowed down and has become flaky. I won't make this mistake again. Meanwhile, any software supplier that proposes to put Spyware on your computer must make it quite clear that it is doing so and give you the option of not installing it. A message along the lines of "do you want to install software that will kill the performance of your computer, send such information about you and your computer as it deems fit to anyone it chooses, and make your computer so unreliable that you will have to re‑install your operating system?" would do nicely.