The average ISP runs almost all Cisco kit - even IOS 12.1(5)T, one of the most recent ones, does not support IPv6, and most ISPs will be running earlier versions. I believe IPv6 support from Cisco (in non-beta IOS versions) will be out this year. I believe Bay and many other routers already support IPv6, but that doesn't matter for most ISPs.
There's also a lot of planning required in how exactly the ISP should switch on IPv6 in their networks - their whole routing setup and DNS infrastructure needs extending, for a start, not to mention the various 6-to-4 conversion schemes. It's not just a matter of turning IPv6 on.
When IPv6 is enabled, you should not have to type 128 bit addresses in any case (did you type in the address for slashdot.org, or did you use DNS).
You say that when costs fall, prices always fall... This is not how the music industry has worked to date - the production cost of CDs is tiny compared to vinyl LPs, but a typical CD costs the same as an LP. So why should they massively reduce costs for online delivery?
Of course, if copy protection works badly enough, these higher prices create an economic incentive for hackers to get round the various mechanisms.
Apple sued Microsoft, unsuccessfully, in the late 1980s for Microsoft's alleged copying of its look and feel. We are lucky this lawsuit failed, otherwise the legal ground for lawsuits against KDE and GNOME for 'copying' Windows and Mac look and feel. The Apple/Microsoft lawsuit was mentioned in the GNU newsletter in 1988
Lotus did something similar, suing companies who copied the command structure of its 1-2-3 spreadsheet. Microsoft, for all its faults, has never pursued look and feel lawsuits.
The Apple and Lotus cases are summarised here - interestingly, Microsoft licensed the Macintosh user interface in 1985, and Apple still sued them.
Are millions of cars really going to be sending their current positions to some massively distributed system (it could never be centralised, of course)? Sounds hugely impractical and very expensive to deploy.
Your Portland bus example is quite irrelevant here - the bus system only has to get the bus's approximate position and then calculate how many route miles are left, and thus get the time to arrival. It doesn't know or care which road the bus is on.
The world has changed - there is now a huge demand for 1U rackmount servers to plug into web farms. Even single websites are using web farms for scalability, of course.
A big issue with web farms is data centre space and cooling - Transmeta can greatly reduce this if it fits many CPUs in a single box within the same heat/power budget. Even if they just have one or two CPUs per box, they can make an impact on this.
Susan Blackmore's excellent book, The Meme Machine, proposes the idea that imitation (of specific actions or behaviours) is at the heart of meme replication. The idea is that you see or hear someone humming a certain tune, and that meme hops neatly into your brain; imitation is the key, i.e. your brain now makes you able to hum the same tune, even if you don't do it straight away. The same arguments apply to art, language, music, and trolling on Slashdot:)
The interesting bit is that her hypothesis has generated testable predictions, including one that specific brain mechanisms would be found that support imitation. It looks like mirror neurons are such a mechanism, supporting her ideas.
Amazon.com has some interesting review comments on this book, see http://www.amazon.com/exec/obidos/ASIN/019286212X/
Windows has been known to crash on VMware running on Linux, but I can assure you VMware does not exit - it just displays the BSOD in the same way a non-virtual PC would.
In some ways, Windows on VMware is actually more stable than Windows on real hardware, largely because VMware emulates hardware that has well proven drivers.
Thanks for the pointer to stow (lhttp://freshmeat.net/appindex/1999/09/23/9381254 46.html) - it ooks very useful for software that comes in.tar.gz files or proprietary installers. I like the way it just adds the minimal set of features necessary, and encodes the installed files as symlinks for easy removal later.
Indeed - right now the only way is a proprietary installer format (e.g. InstallShield's java-based installer, which installs its own JRE as necessary).
EPM, posted earlier, looks pretty good for the Unix world, since it already supports quite a few platforms such as Debian, Red Hat, Solaris, HP-UX and Irix. This is unlikely to ever cover Windows - Microsoft has its own recommended installer format (Windows Installer,.msi) so most packages are migrating to that anyway.
Paraphrased from somewhere: "Never ascribe to malice that which can adequately be explained by incompetence". Not exactly right for this, but close...
I think Intel (and Microsoft, in Whistler-64) are so focused on 64 bit performance that they haven't spent enough time on the 32 bit emulation/translation piece, which will probably be improved in McKinley. Or alternatively, they just don't care about 32 bit performance since IA-64 will largely run on servers, at least initially, and everyone will have produced new server software.
I do think it was hardly worth doing this review with 32 bit tests only - it would have been a lot more useful to put in another hard drive and install Linux IA-64 on that, removing it after the test. At least then they could have compiled some benchmarks from source.
A quick Google search reveals that metacomputing seems to be related to the Grid, i.e. it's a way of gluing disparate systems together (usually supercomputers) with a single set of middleware that makes it much easier to write large-scale distributed applications.
Such systems will be used for all kinds of scientific calculations, as well as telemedicine, distributed virtual reality caves, and so on. The Grid will eventually impinge on us all (e.g. running massive simulations to make a medical diagnosis as you sit in your doctor's surgery).
Depending on the sort of application, the bandwidth demands can be enormous.
You are making this more confusing than it really is, by not using any technical terms that might make sense (e.g. add-drop multiplexer, optical switch). A shame, since your points are valid...
- 'Smoke' - it's hard to work out what you are talking about here - seems like the 'smoke' box is an add-drop multiplexer for DWDM, which puts multiple frequencies (aka wavelengths) from various input fibres on a single output fibre. DWDM is inherently multi-protocol of course, as each wavelength can carry a unique protocol.
- 'Mirrors' - this is just one of the many possible all-optical switching technologies that are under development. These include MEMS (tiny mirrors that can reflect light onto different fibres), electro-holographic Bragg gratings (completely solid state and with useful testing/monitoring features), and even a bizarre technology that involves using inkjet techniques to blow bubbles in and out of place, thereby affecting switching (from Agilent).
The real solution to open ethernet ports is to use authenticated VLANs - require the workstation user to authenticate before being put into a suitable VLAN, and just ignore the MAC address.
A serious answer for once - just buy Checkpoint's Firewall-1 (they popularised stateful inspection). Quite expensive, licensed per host behind the firewall, and not peer reviewed like Linux or BSD firewalling, but pretty popular.
Unfortunately the UK's Mailing Preference Service is not backed by law - any company that is not part of the UK Direct Marketing Association is free to do what it wants. See http://www.dma.org.uk/thedma/cgi-bin/incorporate.p l?&path=../documents/prfro4izm.txt&user=flflflflfl fl
The Data Protection Act does seem to provide some degree of opt-out - every reputable company gives you the chance to opt out of mailings, but you have to hunt to find the tiny box on the form that you need to tick for this. You can't register your name directly with the data protection registrar, only with the DMA's service or individual companies to opt out of their mailings.
I think the IBM 360 architecture was the first really popular range to support 8 bit bytes and 32 bit words - it was also the first time a consistent instruction set architecture was defined over a range of machines of varying size, using microcode. All these features are still used in most CISC processors, such as Pentiums.
Some supporting URLs:
- http://www.dg.com/about/html/ibm_360.html
- http://www.cs.uiowa.edu/~jones/assem/notes/05.html (search for '360')
Re:what about the pdp-8??
on
PDP-10 Revival
·
· Score: 2
Ah, nostalgia... You had floppy disks? I learnt to program on a PDP-8/F, which was the size of two large desktop PC cases and had an ASR-33 teletype for I/O - paper tape or typing/printing only. I was super-impressed when we got a high speed tape reader that could read yards of tape incredibly fast (probably all of hundreds of characters per second).
It had core memory (non-volatile and you could see the bits with a naked eye), like all machines of those days. Normally this held the bootstrap program, which was all of a dozen or so 12-bit instructions, but every now and then it would get corrupted and you'd have to toggle it in by setting 12 switches on the front panel and flipping another switch to say 'deposit'.
Compared to this, the PDP-10 was truly a mainframe - the PDP-8 could run operating systems but ours mainly ran BASIC, which ran on the the bare machine.
I still can't quite work out why anyone is paying for porting GCC to the PDP-10 - I can understand porting GCC just for the sheer hack value, but why would a company pay for this? I don't buy the idea that telcos are still using PDP-10s somewhere - if there are any, the market must be truly tiny and most software would be written in assembler.
I keep doing this, because I prefer Opera most of the time on Linux as well as Windows, and many sites actually prevent use of Opera. I don't mind so much if the site doesn't look great, what's a pain is when someone has taken the trouble to exclude Netscape and Opera from a plain vanilla site.
Amazon.com and its versions in other countries are good arguments for browser independence by the way - no frames, (at least some) ALT tags, and no Javascript, and of course no browser exclusion. And most importantly, they are known by everyone and commercially successful (useful when arguing with PHBs).
However, I've yet to convince clueless sites like www.jamjar.com that they should change their policies, even after stating that I'm not going to be buying a car for them...
Not a T-shirt, but hey... Usenix '83 had a button saying 'Sex, Drugs and UNIX' (yes, Unix was cool once:) - a few years later, another one was produced after the invasion of suits into Unixland, and read 'Condoms, aspirin and POSIX'.
I've found Netscape too buggy on Linux (though strangely, it seems to crash much less, if at all, since I moved from dialup to ADSL...). And Mozilla is horribly slow on some sites, even when using the Back button, on a K6-2/350 with 256 MB of RAM.
Opera, by contrast, is incredibly fast and has yet to crash in its beta 4 incarnation. It's as fast as IE5 on Windows, where I use Opera as well as my main browser.
I'm using Opera 4.0b4 as my main Linux browser already. The only thing stopping me using Opera more is the combination of non-standard sites, particularly those with browser detection, and the Opera team's reluctance to do a 'relax standards conformance' option - some sites display nothing at all, though this is quite rare.
Of course, Opera is commercial on Linux (or ad-supported on Windows if you prefer), but it runs fast even on 486s and has all the features of a modern graphical browser. See http://www.opera.com for a download, and don't forget to report bugs via their online form...
Sites that mandate IE or NS are quite rare, but often come from large companies - for example, jamjar.com (UK car buying site from major insurance company) is unusable even with Netscape on Linux, let alone Opera. The website management there can't or won't take out the line of browser detection code that prevents Linux working, despite several complaints.
You might want to investigate XML, with a suitable DTD, e.g. DocBook for technical manuals. Also, SVG is an XML-based format for vector graphics, which always seemed to be the point at which SGML-based efforts had trouble.
Tool support for this combination may not be so good or inexpensive, but you can be fairly sure the content will survive and be usable in many different environments.
ATM is still around but is increasingly used only within the core of large telcos (to deliver Frame Relay, leased line emulation, and so on, which are still very profitable) and for ADSL (I'm posting this over an ATM 25 Mbps connection from my ADSL router).
IP has its challenges but, when combined with MPLS, can do much of what ATM can do - QoS (with some limitations but much more scalable than ATM), traffic engineering (balancing utilisation of circuits all over the network to meet QoS goals or reduce costs), and closed user groups (aka IP VPNs).
There's absolutely no need to ditch IP for other protocols (except for the upgrade to IPv6 of course:). In particular, the issue of voice packet overhead is addressed through:
- RTP compression (compresses the IP+UDP+RTP header of VoIP packets quite significantly)
- Voice over MPLS (VoMPLS) - puts the RTP payload into an MPLS encapsulation (just 4 bytes of overhead), stripping off the IP and UDP on ingress to the MPLS domain and adding it back on egress.
Also, the much discussed small size of ATM cells is only really of benefit on low speed links - if you have a 1 Gbps link, there is probably no need to pre-empt the transmission of large file transfer packets to squeeze in a VoIP packet ASAP. The IP world has link-level fragmentation techniques (LFI on Cisco's, PPP options on everything else) that mean you get tiny link-level fragments, rather like ATM cells, on low speed links, but avoid the overhead of cells on high speed links.
One way of pushing the intelligence to the edges is to use MPLS to control optical networks - the edge MPLS label-switch router assigns an MPLS label in the form of a wavelength, while the core optical switches know nothing about IP and can perform pure optical switching.
Slashdot Mirror
Only vaguely on topic, but I once saw one of those enormous display screens outside a Las Vegas casino, proudly displaying a BSOD...
The average ISP runs almost all Cisco kit - even IOS 12.1(5)T, one of the most recent ones, does not support IPv6, and most ISPs will be running earlier versions. I believe IPv6 support from Cisco (in non-beta IOS versions) will be out this year. I believe Bay and many other routers already support IPv6, but that doesn't matter for most ISPs.
There's also a lot of planning required in how exactly the ISP should switch on IPv6 in their networks - their whole routing setup and DNS infrastructure needs extending, for a start, not to mention the various 6-to-4 conversion schemes. It's not just a matter of turning IPv6 on.
When IPv6 is enabled, you should not have to type 128 bit addresses in any case (did you type in the address for slashdot.org, or did you use DNS).
You say that when costs fall, prices always fall... This is not how the music industry has worked to date - the production cost of CDs is tiny compared to vinyl LPs, but a typical CD costs the same as an LP. So why should they massively reduce costs for online delivery?
Of course, if copy protection works badly enough, these higher prices create an economic incentive for hackers to get round the various mechanisms.
Apple sued Microsoft, unsuccessfully, in the late 1980s for Microsoft's alleged copying of its look and feel. We are lucky this lawsuit failed, otherwise the legal ground for lawsuits against KDE and GNOME for 'copying' Windows and Mac look and feel. The Apple/Microsoft lawsuit was mentioned in the GNU newsletter in 1988
Lotus did something similar, suing companies who copied the command structure of its 1-2-3 spreadsheet. Microsoft, for all its faults, has never pursued look and feel lawsuits.
The Apple and Lotus cases are summarised here - interestingly, Microsoft licensed the Macintosh user interface in 1985, and Apple still sued them.
Are millions of cars really going to be sending their current positions to some massively distributed system (it could never be centralised, of course)? Sounds hugely impractical and very expensive to deploy.
Your Portland bus example is quite irrelevant here - the bus system only has to get the bus's approximate position and then calculate how many route miles are left, and thus get the time to arrival. It doesn't know or care which road the bus is on.
The world has changed - there is now a huge demand for 1U rackmount servers to plug into web farms. Even single websites are using web farms for scalability, of course.
A big issue with web farms is data centre space and cooling - Transmeta can greatly reduce this if it fits many CPUs in a single box within the same heat/power budget. Even if they just have one or two CPUs per box, they can make an impact on this.
Susan Blackmore's excellent book, The Meme Machine, proposes the idea that imitation (of specific actions or behaviours) is at the heart of meme replication. The idea is that you see or hear someone humming a certain tune, and that meme hops neatly into your brain; imitation is the key, i.e. your brain now makes you able to hum the same tune, even if you don't do it straight away. The same arguments apply to art, language, music, and trolling on Slashdot :)
/
The interesting bit is that her hypothesis has generated testable predictions, including one that specific brain mechanisms would be found that support imitation. It looks like mirror neurons are such a mechanism, supporting her ideas.
Amazon.com has some interesting review comments on this book, see http://www.amazon.com/exec/obidos/ASIN/019286212X
Windows has been known to crash on VMware running on Linux, but I can assure you VMware does not exit - it just displays the BSOD in the same way a non-virtual PC would.
In some ways, Windows on VMware is actually more stable than Windows on real hardware, largely because VMware emulates hardware that has well proven drivers.
Thanks for the pointer to stow (lhttp://freshmeat.net/appindex/1999/09/23/9381254 46.html) - it ooks very useful for software that comes in .tar.gz files or proprietary installers. I like the way it just adds the minimal set of features necessary, and encodes the installed files as symlinks for easy removal later.
Indeed - right now the only way is a proprietary installer format (e.g. InstallShield's java-based installer, which installs its own JRE as necessary).
.msi) so most packages are migrating to that anyway.
EPM, posted earlier, looks pretty good for the Unix world, since it already supports quite a few platforms such as Debian, Red Hat, Solaris, HP-UX and Irix. This is unlikely to ever cover Windows - Microsoft has its own recommended installer format (Windows Installer,
Paraphrased from somewhere: "Never ascribe to malice that which can adequately be explained by incompetence". Not exactly right for this, but close...
I think Intel (and Microsoft, in Whistler-64) are so focused on 64 bit performance that they haven't spent enough time on the 32 bit emulation/translation piece, which will probably be improved in McKinley. Or alternatively, they just don't care about 32 bit performance since IA-64 will largely run on servers, at least initially, and everyone will have produced new server software.
I do think it was hardly worth doing this review with 32 bit tests only - it would have been a lot more useful to put in another hard drive and install Linux IA-64 on that, removing it after the test. At least then they could have compiled some benchmarks from source.
A quick Google search reveals that metacomputing seems to be related to the Grid, i.e. it's a way of gluing disparate systems together (usually supercomputers) with a single set of middleware that makes it much easier to write large-scale distributed applications.
Such systems will be used for all kinds of scientific calculations, as well as telemedicine, distributed virtual reality caves, and so on. The Grid will eventually impinge on us all (e.g. running massive simulations to make a medical diagnosis as you sit in your doctor's surgery).
Depending on the sort of application, the bandwidth demands can be enormous.
You are making this more confusing than it really is, by not using any technical terms that might make sense (e.g. add-drop multiplexer, optical switch). A shame, since your points are valid...
- 'Smoke' - it's hard to work out what you are talking about here - seems like the 'smoke' box is an add-drop multiplexer for DWDM, which puts multiple frequencies (aka wavelengths) from various input fibres on a single output fibre. DWDM is inherently multi-protocol of course, as each wavelength can carry a unique protocol.
- 'Mirrors' - this is just one of the many possible all-optical switching technologies that are under development. These include MEMS (tiny mirrors that can reflect light onto different fibres), electro-holographic Bragg gratings (completely solid state and with useful testing/monitoring features), and even a bizarre technology that involves using inkjet techniques to blow bubbles in and out of place, thereby affecting switching (from Agilent).
The real solution to open ethernet ports is to use authenticated VLANs - require the workstation user to authenticate before being put into a suitable VLAN, and just ignore the MAC address.
A serious answer for once - just buy Checkpoint's Firewall-1 (they popularised stateful inspection). Quite expensive, licensed per host behind the firewall, and not peer reviewed like Linux or BSD firewalling, but pretty popular.
Unfortunately the UK's Mailing Preference Service is not backed by law - any company that is not part of the UK Direct Marketing Association is free to do what it wants. See http://www.dma.org.uk/thedma/cgi-bin/incorporate.p l?&path=../documents/prfro4izm.txt&user=flflflflfl fl
The Data Protection Act does seem to provide some degree of opt-out - every reputable company gives you the chance to opt out of mailings, but you have to hunt to find the tiny box on the form that you need to tick for this. You can't register your name directly with the data protection registrar, only with the DMA's service or individual companies to opt out of their mailings.
I think the IBM 360 architecture was the first really popular range to support 8 bit bytes and 32 bit words - it was also the first time a consistent instruction set architecture was defined over a range of machines of varying size, using microcode. All these features are still used in most CISC processors, such as Pentiums.
l (search for '360')
Some supporting URLs:
- http://www.dg.com/about/html/ibm_360.html
- http://www.cs.uiowa.edu/~jones/assem/notes/05.htm
Ah, nostalgia... You had floppy disks? I learnt to program on a PDP-8/F, which was the size of two large desktop PC cases and had an ASR-33 teletype for I/O - paper tape or typing/printing only. I was super-impressed when we got a high speed tape reader that could read yards of tape incredibly fast (probably all of hundreds of characters per second).
It had core memory (non-volatile and you could see the bits with a naked eye), like all machines of those days. Normally this held the bootstrap program, which was all of a dozen or so 12-bit instructions, but every now and then it would get corrupted and you'd have to toggle it in by setting 12 switches on the front panel and flipping another switch to say 'deposit'.
Compared to this, the PDP-10 was truly a mainframe - the PDP-8 could run operating systems but ours mainly ran BASIC, which ran on the the bare machine.
I still can't quite work out why anyone is paying for porting GCC to the PDP-10 - I can understand porting GCC just for the sheer hack value, but why would a company pay for this? I don't buy the idea that telcos are still using PDP-10s somewhere - if there are any, the market must be truly tiny and most software would be written in assembler.
I keep doing this, because I prefer Opera most of the time on Linux as well as Windows, and many sites actually prevent use of Opera. I don't mind so much if the site doesn't look great, what's a pain is when someone has taken the trouble to exclude Netscape and Opera from a plain vanilla site.
Amazon.com and its versions in other countries are good arguments for browser independence by the way - no frames, (at least some) ALT tags, and no Javascript, and of course no browser exclusion. And most importantly, they are known by everyone and commercially successful (useful when arguing with PHBs).
However, I've yet to convince clueless sites like www.jamjar.com that they should change their policies, even after stating that I'm not going to be buying a car for them...
Not a T-shirt, but hey... Usenix '83 had a button saying 'Sex, Drugs and UNIX' (yes, Unix was cool once :) - a few years later, another one was produced after the invasion of suits into Unixland, and read 'Condoms, aspirin and POSIX'.
I've found Netscape too buggy on Linux (though strangely, it seems to crash much less, if at all, since I moved from dialup to ADSL...). And Mozilla is horribly slow on some sites, even when using the Back button, on a K6-2/350 with 256 MB of RAM.
Opera, by contrast, is incredibly fast and has yet to crash in its beta 4 incarnation. It's as fast as IE5 on Windows, where I use Opera as well as my main browser.
I'm using Opera 4.0b4 as my main Linux browser already. The only thing stopping me using Opera more is the combination of non-standard sites, particularly those with browser detection, and the Opera team's reluctance to do a 'relax standards conformance' option - some sites display nothing at all, though this is quite rare.
Of course, Opera is commercial on Linux (or ad-supported on Windows if you prefer), but it runs fast even on 486s and has all the features of a modern graphical browser. See http://www.opera.com for a download, and don't forget to report bugs via their online form...
Sites that mandate IE or NS are quite rare, but often come from large companies - for example, jamjar.com (UK car buying site from major insurance company) is unusable even with Netscape on Linux, let alone Opera. The website management there can't or won't take out the line of browser detection code that prevents Linux working, despite several complaints.
You might want to investigate XML, with a suitable DTD, e.g. DocBook for technical manuals. Also, SVG is an XML-based format for vector graphics, which always seemed to be the point at which SGML-based efforts had trouble.
Tool support for this combination may not be so good or inexpensive, but you can be fairly sure the content will survive and be usable in many different environments.
ATM is still around but is increasingly used only within the core of large telcos (to deliver Frame Relay, leased line emulation, and so on, which are still very profitable) and for ADSL (I'm posting this over an ATM 25 Mbps connection from my ADSL router).
:). In particular, the issue of voice packet overhead is addressed through:
IP has its challenges but, when combined with MPLS, can do much of what ATM can do - QoS (with some limitations but much more scalable than ATM), traffic engineering (balancing utilisation of circuits all over the network to meet QoS goals or reduce costs), and closed user groups (aka IP VPNs).
There's absolutely no need to ditch IP for other protocols (except for the upgrade to IPv6 of course
- RTP compression (compresses the IP+UDP+RTP header of VoIP packets quite significantly)
- Voice over MPLS (VoMPLS) - puts the RTP payload into an MPLS encapsulation (just 4 bytes of overhead), stripping off the IP and UDP on ingress to the MPLS domain and adding it back on egress.
Also, the much discussed small size of ATM cells is only really of benefit on low speed links - if you have a 1 Gbps link, there is probably no need to pre-empt the transmission of large file transfer packets to squeeze in a VoIP packet ASAP. The IP world has link-level fragmentation techniques (LFI on Cisco's, PPP options on everything else) that mean you get tiny link-level fragments, rather like ATM cells, on low speed links, but avoid the overhead of cells on high speed links.
One way of pushing the intelligence to the edges is to use MPLS to control optical networks - the edge MPLS label-switch router assigns an MPLS label in the form of a wavelength, while the core optical switches know nothing about IP and can perform pure optical switching.