Find out about NAT-PT (NAT-Protocol Translation) - this takes IPv6 packets and NATs them, translating to IPv4 at the same time. Useful to talk to the IPv4 world from your domain.
To talk to other IPv6 domains, investigate the various tunnelling approaches for 6 over 4. In particular, there's one that automatically builds tunnels as required (6to4, I think).
Actually you can get QoS with plain IPv4 (or rather CoS, class of service, which is good enough for many situations and much easier to deploy).
MPLS adds traffic engineering, which is loosely the ability to create virtual circuits a bit like ATM - so you can get harder QoS and load-balance traffic across the network, onto underutilised links.
Both are independent of IPv6, and IPSec is also a red herring since it works fine in IPv4.
IPv6 may not be the perfect approach, but IMO it has real benefits in autoconfiguration (renumber from a single point and watch the changes just happen) and many other areas - unfortunately these don't make good soundbites.
Probably the most important issue is getting rid of NAT - though this may be optimistic since many organisations are very happy with their proxy-based firewalls, even worse for interoperability than NAT, and won't change them in a hurry.
IPv6 includes a variety of MAC addresses within its address format (though the embedded MAC part can also be set to whatever you want).
The main reason to go to 128 bit addresses is to have a single unique IP address across all the billions (literally) of IP-enabled phones, web tablets, Internet appliances, and so on.
DBMSs such as Oracle also gather multiple transactions (each a set of writes) and do a 'group commit' - the transactions are committed atomically by writing to a journal.
The point is that you can group writes based on transactions and performance, or based purely on performance - Oracle and some journalling filesystems do the former, Linux and others do the latter.
In both cases you end up with a long sequential write to the journal file - certainly Oracle claimed a big speedup in transactions per second when this was introduced.
Remember that the Visor runs a CISC processor (Dragonball EZ, basically a 680x0 based microcontroller), which is very efficient in code size. Also, Palm apps are well known for being efficient in use of data space.
I have at least 30 third party apps, including a web browser with 1 MB cache, in a 4MB Palm IIIx. I've just upgraded to a Visor with 8MB, but that's largely for web browsing and to read books on the thing. If you just want to use it as an organiser, even 2MB is overkill.
Having said that, the iPaq is tempting because it can run Linux - although there are probably few useful apps for Linux at present, the stability is attractive (my Palm III just had to be hard-reset, losing all data - first time in months, but still...)
Good luck getting a week's battery life with a hard drive, colour TFT screen, and 500+ MHz Intel-compatible processor. Oh, the Psion doesn't have any of those, does it... Not surprising it has a completely different level of power usage, then.
Try comparing apples to apples - it sounds like you would be better off using a WinCE notebook-format device, maybe with Linux on it since this is slashdot:)
Why on earth was that comment moderated as Flamebait? Seemed like a reasonable set of proposals and well-thought-out. Moderators should not be marking something as flamebait just because they disagree with it...
Interesting - some more details are at http://www.adabas.com/overv11.htm - as well as Oracle compatibility, it has Access database migration as one interesting feature, along with JDBC, ODBC, Perl DBI, Tcl/Tk support, and so on.
I wasn't aware that SAP had acquired Adabas-D, but it looks quite useful - certainly quite fully featured.
Unfortunately NTOP only shows hosts, not TCP/UDP port numbers, in its breakdown, so it's not very useful for this application.
MRTG is a nice tool, but limited by what SNMP/RMON2 MIBs are available in typical routers and switches - by far the majority of routers and most low to mid-end switches do not support RMON2, which is the only SNMP MIB that would let you monitor bandwidth by TCP/UDP port.
IPaudit or IPmeter are the best open source options I have found; IPaudit can just sniff the network like ntop.
If you have Cisco routers, you can enable something called NetFlow in IOS 12.0 or higher, which does this sort of accounting - search for cflowd, which is an open source data collection tool. But be careful, NetFlow can easily overload a Cisco router's CPU, and you need to have admin access to your Cisco router anyway.
If you have a spare PC that can have Linux installed and be connected to the ethernet segment serving your Internet access router, something like ipaudit may be enough. It can monitor TCP/UDP ports, which you can't do with ipac.
If you have an ethernet switch serving the access router (quite likely), you will need to set up the switch to 'span' or 'mirror' the port serving the router to another port (serving your monitoring box). This just replicates (broadcasts) the traffic seen in and outbound on the access router's switch port, into the monitoring port.
Alternatively you could put a hub between the switch and the router, but your network manager is unlikely to be happy about this.
Commercial tools to do this are astonishingly expensive, by the way - there are things called RMON2 probes that do more or less what ipacct does, but with more features and SNMP-accessible MIBs for the results. Bandwidth management boxes such as Packeteer do something similar, but these are also quite expensive.
If any entrepreneur out there feels like doing embedded Linux or BSD boxes that monitor and maybe shape traffic (Linux's queuing features in 2.2 or later are very comprehensive), you would have quite a market. Even more so if you worked on cish (an open source emulation of the Cisco command line interface) so that standard QoS management tools could configure your box just like a Cisco router. Another useful standard to look at is RTFM (real-time traffic flow management) from the IETF, implemented by Netramet, which is supported by ipmeter.
Of course, I have an ulterior motive;) My company (Orchestream) does QoS-management tools, as do others, which talk to Cisco routers. However, it's tough to justify modifying our product to talk to the Linux QoS stuff until there's enough volume of Linux routers out there. I would like to see us configuring Linux QoS-enabled routers, though.
"I'd sure love to see a multicast that could handle bandwidth overloads, but I don't think it exists yet. "
There is something called layered encoding that works well with multicast, but only really for audio/video rather than gaming/DIS: you send N streams of data for a single input multimedia stream, each on a differerent multicast group. Clients tune in to streams 1 to M, where M = N - stream 1 gives OK quality, adding stream 2 improves quality, stream 3 improves it further, and so on.
The result is that clients on poor connections can reduce the amount of detail. This might work for DIS if you could come up with a 'high detail' type channel that provides extra info for high-end clients on broadband connections. However, the filtering you were talking about it too hard for this sort of mechanism, so server-based filtering is probably still the way to go.
In the longer term, it's conceivable that routers may be able to run downloadable code, or even be programmed by the packets on the fly, which would enable this - however, this is very speculative. Search for tanenbaum and active networks for more info.
Interesting that the classic DIS stuff is so primitive, but these are application issues rather than network ones, and it will be potentially hard to scale up the MMP servers if millions of people start playing these games.
If you are using multicast (which I think modern DISs are aimed at, certainly the IETF LSMA work is) it's crucial to intelligently design the application and make good use of the multicast groups (don't put everyone in one big group, for example).
Some mix of MMP and multicast may be useful - e.g. multicast between servers or to broadband clients, and special protocols to other clients. The big advantage of multicast is that it's just a routing/forwarding technology, so if your client has the bandwidth it's more efficient to use multicast groups where you can.
There are many ways of bypassing NTFS security if you have physical access to the disk - Linux NTFS is one, but there are commercial tools as well that let you use NTFS on DOS or Win9x.
Encrypting files or filesystems is the only way to guard against physical access to the disk.
For a view of what might be needed from the network to handle this sort of application, see http://www.ietf.org/html.charters/lsma-charter.htm l - this is an IETF working group that has been documenting and classifying the types of network demands (stringent QoS needs, very large multicast groups, etc) made by distributed interactive simulations (DISs).
DISs can be networked military simulators located around the world (this is how the US Army rehearsed the Gulf War and continues to train its tank crew), or massively multiplayer games.
Such simulations/games may in the future have up to 100,000 simulated objects, representing human or computer-managed players.
You are overreacting just a touch.... The definite article is fairly obvious from the context - why waste screen space and user time on 'Shut Down *The* Computer', when it's clear that you are not shutting down any other computer?
Or maybe there should be 'Shutdown A Computer', Shutdown Any Computer You Feel Like', etc?
One thing MS has done right in Whistler is include a personal firewall (i.e. intended for desktops or laptops that are directly Net connected).
This is important for the whole Net, in order to reduce the number of DoS attacks launched from compromised machines, particularly as cable modems and ADSL become more common.
I hope Linux distros and KDE/Gnome are going to include similar features - I know that ipchains is there by default, but what I want to see is that the basic firewall is installed as part of the install process, as a result of asking 'will this machine always be used behind a firewall?'
The Windows world is getting into personal firewalls - McAfee (formerly Conseal), Norton and the idiosyncratic ZoneGuard are some examples.
Probably MS were going to do this anyway, but I remember talking to a fairly senior program manager a year or two ago, who was on the Win2000 team, and saying how important it was to have personal firewalls, so maybe I can claim some credit:) Amusingly, he wrote the details down in his Palm III...
"The main reason to have a case-sensitive filesystem is to support the generation of randomly named temp files, such as 'GBVhX88r' and 'gbVHx88R'."
This is hardly a reason - simply lowercase all generated temp file names before creating the files, or don't create temp names with upper case letters.
Actually, AOL has ported their normal AIM client to Linux - it's available in beta, and works OK though I reverted to Gaim for reasons I can't remember any more...
Yahoo Messenger is also available on Linux as well - quite weird having these official versions coming out...
This is not a very elegant solution, but it worked for me... Try updating to the latest mozilla nightly build, and then installing Helix Gnome. Galeon is now working very nicely, and is very fast. Still a bit primitive, but not bad.
This is common in most decent laptops, and is sometimes called Hibernate - it requires extra code to reset devices to the state they were in at hibernate time, but it always worked very well on my IBM Thinkpad under Windows.
To get the best of both worlds, it would be very handy if you could use logical volume management (or maybe something simpler) to ensure that the/tmp filesystem started out using RAM disk, then migrated onto a second physical volume on disk.
But perhaps the real issue is filesystem-specific caching parameters - if you could configure the/tmp filesystem caching to be much more aggressive, using more memory, this would be self-tuning (i.e. expanding to disk when needed) and probably work better all round. Though perhaps a special filesystem would still be needed to avoid writing to disk unless you've run out of RAM.
Patents are not granted for products - they are granted for innovative techniques, methods or processes for solving a problem.
The one-click method is not so different to being billed for premium information on a phone call, e.g. a weather forecast. The phone company has all your billing info, so it only takes one action (making the call) to get the goods (weather forecast, delivered as audio) and be billed at the same time.
Your list of functions is useful but it includes a great mixture of different types of things, and it's worth breaking them out:
- some that depend on Web-based, collaborative package reviews (which don't really exist yet and IMO are a big need for open source). This needs addressing, ideally using websites that have proper XML tagging so that programs can extract the reviews and search/analyse them more easily.
- some that depend on the package (e.g. help files or intro docs, and the ability to browse package docs before installing).
- some which are true package management issues, e.g. don't install packages that would require upgrades to libraries already in use. This is an example of policy - it would be helpful if a standard approach to such policies could be defined, then the various packaging tools could make sure they support this, and the GUIs could make it easy to specify this.
I think a lot of these issues are being addressed, but in a piecemeal fashion, e.g. Freshmeat.net is great on listing packages but not on reviews, various packaging GUIs may make it possible to more easily browse docs and specify more complex policies.
It might be useful to have a packaging framework initiative that tries to encourage various efforts in these areas and acts as a central point of information and even standards (e.g. standardised XML tags for reviews).
My main issue with packaging tools is that even with GUIs they require a lot of user expertise - first of all, where to find the RPM, then checking it's the latest version and compatible with your system, then which mirror to select for a fast download (a separate problem but one that should be automated, see the SPAND project).
Then there's the issue of managing or archiving the downloaded packages once installed. It would be useful if there was a log file of all installed packages + how they were installed, held in the archive directory, so that you could just back up this directory to get a fairly quick and dirty restore of this system (or to ease mirroring installs on other systems).
Slashdot Mirror
Find out about NAT-PT (NAT-Protocol Translation) - this takes IPv6 packets and NATs them, translating to IPv4 at the same time. Useful to talk to the IPv4 world from your domain.
To talk to other IPv6 domains, investigate the various tunnelling approaches for 6 over 4. In particular, there's one that automatically builds tunnels as required (6to4, I think).
Actually you can get QoS with plain IPv4 (or rather CoS, class of service, which is good enough for many situations and much easier to deploy).
MPLS adds traffic engineering, which is loosely the ability to create virtual circuits a bit like ATM - so you can get harder QoS and load-balance traffic across the network, onto underutilised links.
Both are independent of IPv6, and IPSec is also a red herring since it works fine in IPv4.
IPv6 may not be the perfect approach, but IMO it has real benefits in autoconfiguration (renumber from a single point and watch the changes just happen) and many other areas - unfortunately these don't make good soundbites.
Probably the most important issue is getting rid of NAT - though this may be optimistic since many organisations are very happy with their proxy-based firewalls, even worse for interoperability than NAT, and won't change them in a hurry.
IPv6 includes a variety of MAC addresses within its address format (though the embedded MAC part can also be set to whatever you want).
The main reason to go to 128 bit addresses is to have a single unique IP address across all the billions (literally) of IP-enabled phones, web tablets, Internet appliances, and so on.
DBMSs such as Oracle also gather multiple transactions (each a set of writes) and do a 'group commit' - the transactions are committed atomically by writing to a journal.
The point is that you can group writes based on transactions and performance, or based purely on performance - Oracle and some journalling filesystems do the former, Linux and others do the latter.
In both cases you end up with a long sequential write to the journal file - certainly Oracle claimed a big speedup in transactions per second when this was introduced.
Remember that the Visor runs a CISC processor (Dragonball EZ, basically a 680x0 based microcontroller), which is very efficient in code size. Also, Palm apps are well known for being efficient in use of data space.
I have at least 30 third party apps, including a web browser with 1 MB cache, in a 4MB Palm IIIx. I've just upgraded to a Visor with 8MB, but that's largely for web browsing and to read books on the thing. If you just want to use it as an organiser, even 2MB is overkill.
Having said that, the iPaq is tempting because it can run Linux - although there are probably few useful apps for Linux at present, the stability is attractive (my Palm III just had to be hard-reset, losing all data - first time in months, but still...)
Good luck getting a week's battery life with a hard drive, colour TFT screen, and 500+ MHz Intel-compatible processor. Oh, the Psion doesn't have any of those, does it... Not surprising it has a completely different level of power usage, then.
:)
Try comparing apples to apples - it sounds like you would be better off using a WinCE notebook-format device, maybe with Linux on it since this is slashdot
Why on earth was that comment moderated as Flamebait? Seemed like a reasonable set of proposals and well-thought-out. Moderators should not be marking something as flamebait just because they disagree with it...
Interesting - some more details are at http://www.adabas.com/overv11.htm - as well as Oracle compatibility, it has Access database migration as one interesting feature, along with JDBC, ODBC, Perl DBI, Tcl/Tk support, and so on.
I wasn't aware that SAP had acquired Adabas-D, but it looks quite useful - certainly quite fully featured.
Unfortunately NTOP only shows hosts, not TCP/UDP port numbers, in its breakdown, so it's not very useful for this application.
MRTG is a nice tool, but limited by what SNMP/RMON2 MIBs are available in typical routers and switches - by far the majority of routers and most low to mid-end switches do not support RMON2, which is the only SNMP MIB that would let you monitor bandwidth by TCP/UDP port.
IPaudit or IPmeter are the best open source options I have found; IPaudit can just sniff the network like ntop.
If you have Cisco routers, you can enable something called NetFlow in IOS 12.0 or higher, which does this sort of accounting - search for cflowd, which is an open source data collection tool. But be careful, NetFlow can easily overload a Cisco router's CPU, and you need to have admin access to your Cisco router anyway.
;) My company (Orchestream) does QoS-management tools, as do others, which talk to Cisco routers. However, it's tough to justify modifying our product to talk to the Linux QoS stuff until there's enough volume of Linux routers out there. I would like to see us configuring Linux QoS-enabled routers, though.
i des/97-indenst97-diplom/stuttg97/tsld001 .htm
If you have a spare PC that can have Linux installed and be connected to the ethernet segment serving your Internet access router, something like ipaudit may be enough. It can monitor TCP/UDP ports, which you can't do with ipac.
If you have an ethernet switch serving the access router (quite likely), you will need to set up the switch to 'span' or 'mirror' the port serving the router to another port (serving your monitoring box). This just replicates (broadcasts) the traffic seen in and outbound on the access router's switch port, into the monitoring port.
Alternatively you could put a hub between the switch and the router, but your network manager is unlikely to be happy about this.
Commercial tools to do this are astonishingly expensive, by the way - there are things called RMON2 probes that do more or less what ipacct does, but with more features and SNMP-accessible MIBs for the results. Bandwidth management boxes such as Packeteer do something similar, but these are also quite expensive.
If any entrepreneur out there feels like doing embedded Linux or BSD boxes that monitor and maybe shape traffic (Linux's queuing features in 2.2 or later are very comprehensive), you would have quite a market. Even more so if you worked on cish (an open source emulation of the Cisco command line interface) so that standard QoS management tools could configure your box just like a Cisco router. Another useful standard to look at is RTFM (real-time traffic flow management) from the IETF, implemented by Netramet, which is supported by ipmeter.
Of course, I have an ulterior motive
Some useful links:
- cish - http://freshmeat.net/projects/cish/
- IPaudit - http://freshmeat.net/projects/ipaudit/
- IPmeter - http://www.ipmeter.com/
- NetraMet - http://www.auckland.ac.nz/net/NeTraMet/
- RTFM and other tools - http://www.mathematik.uni-stuttgart.de/~floeff/sl
- RTFM home - http://www.auckland.ac.nz/net/Internet/rtfm/
"I'd sure love to see a multicast that could handle bandwidth overloads, but I don't think it exists yet. "
There is something called layered encoding that works well with multicast, but only really for audio/video rather than gaming/DIS: you send N streams of data for a single input multimedia stream, each on a differerent multicast group. Clients tune in to streams 1 to M, where M = N - stream 1 gives OK quality, adding stream 2 improves quality, stream 3 improves it further, and so on.
The result is that clients on poor connections can reduce the amount of detail. This might work for DIS if you could come up with a 'high detail' type channel that provides extra info for high-end clients on broadband connections. However, the filtering you were talking about it too hard for this sort of mechanism, so server-based filtering is probably still the way to go.
In the longer term, it's conceivable that routers may be able to run downloadable code, or even be programmed by the packets on the fly, which would enable this - however, this is very speculative. Search for tanenbaum and active networks for more info.
Interesting that the classic DIS stuff is so primitive, but these are application issues rather than network ones, and it will be potentially hard to scale up the MMP servers if millions of people start playing these games.
If you are using multicast (which I think modern DISs are aimed at, certainly the IETF LSMA work is) it's crucial to intelligently design the application and make good use of the multicast groups (don't put everyone in one big group, for example).
Some mix of MMP and multicast may be useful - e.g. multicast between servers or to broadband clients, and special protocols to other clients. The big advantage of multicast is that it's just a routing/forwarding technology, so if your client has the bandwidth it's more efficient to use multicast groups where you can.
There are many ways of bypassing NTFS security if you have physical access to the disk - Linux NTFS is one, but there are commercial tools as well that let you use NTFS on DOS or Win9x.
Encrypting files or filesystems is the only way to guard against physical access to the disk.
For a view of what might be needed from the network to handle this sort of application, see http://www.ietf.org/html.charters/lsma-charter.htm l - this is an IETF working group that has been documenting and classifying the types of network demands (stringent QoS needs, very large multicast groups, etc) made by distributed interactive simulations (DISs).
DISs can be networked military simulators located around the world (this is how the US Army rehearsed the Gulf War and continues to train its tank crew), or massively multiplayer games.
Such simulations/games may in the future have up to 100,000 simulated objects, representing human or computer-managed players.
You are overreacting just a touch.... The definite article is fairly obvious from the context - why waste screen space and user time on 'Shut Down *The* Computer', when it's clear that you are not shutting down any other computer?
Or maybe there should be 'Shutdown A Computer', Shutdown Any Computer You Feel Like', etc?
One thing MS has done right in Whistler is include a personal firewall (i.e. intended for desktops or laptops that are directly Net connected).
:) Amusingly, he wrote the details down in his Palm III...
This is important for the whole Net, in order to reduce the number of DoS attacks launched from compromised machines, particularly as cable modems and ADSL become more common.
I hope Linux distros and KDE/Gnome are going to include similar features - I know that ipchains is there by default, but what I want to see is that the basic firewall is installed as part of the install process, as a result of asking 'will this machine always be used behind a firewall?'
The Windows world is getting into personal firewalls - McAfee (formerly Conseal), Norton and the idiosyncratic ZoneGuard are some examples.
Probably MS were going to do this anyway, but I remember talking to a fairly senior program manager a year or two ago, who was on the Win2000 team, and saying how important it was to have personal firewalls, so maybe I can claim some credit
"The main reason to have a case-sensitive filesystem is to support the generation of randomly named temp files, such as 'GBVhX88r' and 'gbVHx88R'."
This is hardly a reason - simply lowercase all generated temp file names before creating the files, or don't create temp names with upper case letters.
Actually, AOL has ported their normal AIM client to Linux - it's available in beta, and works OK though I reverted to Gaim for reasons I can't remember any more...
Yahoo Messenger is also available on Linux as well - quite weird having these official versions coming out...
Somebody moderate up the message I'm commenting to, it actually has a solution...
This is not a very elegant solution, but it worked for me... Try updating to the latest mozilla nightly build, and then installing Helix Gnome. Galeon is now working very nicely, and is very fast. Still a bit primitive, but not bad.
Trust me, bad software does exist :) The main reason that choice is so important is to enable users to choose *good* software...
This is common in most decent laptops, and is sometimes called Hibernate - it requires extra code to reset devices to the state they were in at hibernate time, but it always worked very well on my IBM Thinkpad under Windows.
To get the best of both worlds, it would be very handy if you could use logical volume management (or maybe something simpler) to ensure that the /tmp filesystem started out using RAM disk, then migrated onto a second physical volume on disk.
/tmp filesystem caching to be much more aggressive, using more memory, this would be self-tuning (i.e. expanding to disk when needed) and probably work better all round. Though perhaps a special filesystem would still be needed to avoid writing to disk unless you've run out of RAM.
But perhaps the real issue is filesystem-specific caching parameters - if you could configure the
Patents are not granted for products - they are granted for innovative techniques, methods or processes for solving a problem.
The one-click method is not so different to being billed for premium information on a phone call, e.g. a weather forecast. The phone company has all your billing info, so it only takes one action (making the call) to get the goods (weather forecast, delivered as audio) and be billed at the same time.
Your list of functions is useful but it includes a great mixture of different types of things, and it's worth breaking them out:
- some that depend on Web-based, collaborative package reviews (which don't really exist yet and IMO are a big need for open source). This needs addressing, ideally using websites that have proper XML tagging so that programs can extract the reviews and search/analyse them more easily.
- some that depend on the package (e.g. help files or intro docs, and the ability to browse package docs before installing).
- some which are true package management issues, e.g. don't install packages that would require upgrades to libraries already in use. This is an example of policy - it would be helpful if a standard approach to such policies could be defined, then the various packaging tools could make sure they support this, and the GUIs could make it easy to specify this.
I think a lot of these issues are being addressed, but in a piecemeal fashion, e.g. Freshmeat.net is great on listing packages but not on reviews, various packaging GUIs may make it possible to more easily browse docs and specify more complex policies.
It might be useful to have a packaging framework initiative that tries to encourage various efforts in these areas and acts as a central point of information and even standards (e.g. standardised XML tags for reviews).
My main issue with packaging tools is that even with GUIs they require a lot of user expertise - first of all, where to find the RPM, then checking it's the latest version and compatible with your system, then which mirror to select for a fast download (a separate problem but one that should be automated, see the SPAND project).
Then there's the issue of managing or archiving the downloaded packages once installed. It would be useful if there was a log file of all installed packages + how they were installed, held in the archive directory, so that you could just back up this directory to get a fairly quick and dirty restore of this system (or to ease mirroring installs on other systems).