I think softupdates background fsck uses snapshots (like NetApp has), which are *very* cool. They let you take a consistent immediate backup. (Not in 4.4, but something to look forward to...)
If you have windows boxen proxied through a squid (or, behind a unix router where you could install a squid as a transparent proxy) then you can do this:
Obviously it quite an easy filter to come up with, but I may as well post it for anyone that didn't think of it. Bit easier than reconfiguring 4 gazillion IE boxen and fielding all the calls about websites needing VBS/Javascript not working after you've fixed people's machines.
Running the mail through an ssh tunnel is a good and secure way to do things, since you can use host keys to ensure that your mail is only sent to your box and not the box of someone who just picked up your old IP address.
http://www.postfix.org/ current versions auto-configure spam protection suitably for most users (allows subnet to relay but nobody else). Of course you can change it how you want, but out-the-box defaults are fantastic.
The message body really isn't much use. Read the article - "information from [...] user and connection logs." I can't imagine anyone with that kind of userbase logging more than To+From+Timestamp+IP from email unless they know beforehand what to look for. I'm sure it's difficult enough keeping the damn thing running as it is! Same with IM - no point looking at content, but IP addresses, combined with Calling-Station-Id from the RADIUS logs, you have a lot of information to feed into some correlation software, that's a lot more interesting.
Trouble is, as things stand they do interfere, a lot of work has been done on this. FHSS (including 802.11 FH devices) can ruin the performance of a well-tuned DHSS running in the band. IEEE 802.15 TG2 is working on mitigating this in Bluetooth by firmware changes to existing/new kit to reduce likelihood of both transmitting at the same time. Plus in the US, FCC has recently approved changes to band usage that will allow FHSS (incl. Bluetooth) to be more adaptive. Looks like there may be a path to allow OFDM on 2.4GHz soon too (I'd be interested to know if anyone has read anything on OFDM/BT interop? I guess it stands better chance than DSSS since it's multiple carrier - maybe there is still a chance for 802.11g...)
Start here if you want to know more...let's just hope that any modifications to Bluetooth required are done now before many mobile phones are sold. It is next to impossible to get firmware updated on non-tech consumer kit once it's out in the wild (unless there are other, show-stopper, bugs which stop the phone from working correctly - ha, like that's never happened before has it Nokia...;)
Funny really, I'd have thought Bluetooth would have been better suited to 5GHz. The higher bandwidth available would be useful (both to increase speed and keep bursts short), and the lower range that's possible isn't such a problem for the intended use of Bluetooth.
Mind you, it's not all bad - personally I'd be quite a bit more interested to see this kind of tech used for the backbone - a neighbourhood is much easier to wire than a long distance link... to get this kind of DX you'll be using fairly tight beams, much less subject to interference, and as long as Bluetooth can be made to tread a little less heavily over local 802.11b base-station installations than it has in the past and share the band fairly then the current generation of cheap(ish) WLAN kit will still be useful and there will be much fun to be had by all...
I hope it does get sorted, things like this could be very cool and while I really don't think d*mned wireless mobile phone headsets are worth the use of free spectrum (*why* can't this kind of thing be run out of *licensed* band if it's run by a commercial service! there should be plenty of that available for lowpower...) I live in hope that everyone'll be able to live together, not stomp all over each other's signal and be happy...
For all but opening new windows, Mozilla is faster than IE on my windows box. Even with the recently added slow-down-turbo-mode-to-accommodate-the-very-few-p eople-who-use-profiles code it's still faster to start.
The Proxomitron is a great adblocker for Windows - lots more modifications possible than with Junkbuster, lots of blocking for certain types of javascript etc., if you want heavy modifications this is definitely the most flexible software.
Junkbuster isn't too good at HTTP protocol, you have to fallback to 1.0 which slows things down a lot (see bugzilla).
Paper motherboards, plastic cases? I guess you've only been looking at the low-end NetServers. Some of the rackmounts are lovely, and certainly in the UK they're a lot easier to buy than similarly specced Proliant boxes..
PPP over SSH isn't a very good solution when there's any packet loss (a definite possibility with wireless) since the two layers of TCP interfere. cipe is probably worth a look (NT and Linux).
300 - that was back when modems only transmitted 1 bit per baud, so I guess you can be forgiven...
Hey, perhaps that's it - maybe they're talking about 28.8kbaud:-)
Slashdot Mirror
I think softupdates background fsck uses snapshots (like NetApp has), which are *very* cool. They let you take a consistent immediate backup. (Not in 4.4, but something to look forward to...)
acl umbricus_microsoftius url_regex \.eml$
http_access deny umbricus_microsoftius
Obviously it quite an easy filter to come up with, but I may as well post it for anyone that didn't think of it. Bit easier than reconfiguring 4 gazillion IE boxen and fielding all the calls about websites needing VBS/Javascript not working after you've fixed people's machines.
http://www.postfix.org/ current versions auto-configure spam protection suitably for most users (allows subnet to relay but nobody else). Of course you can change it how you want, but out-the-box defaults are fantastic.
And this is really all they're cutting in the US?
The message body really isn't much use. Read the article - "information from [...] user and connection logs." I can't imagine anyone with that kind of userbase logging more than To+From+Timestamp+IP from email unless they know beforehand what to look for. I'm sure it's difficult enough keeping the damn thing running as it is! Same with IM - no point looking at content, but IP addresses, combined with Calling-Station-Id from the RADIUS logs, you have a lot of information to feed into some correlation software, that's a lot more interesting.
Trouble is, as things stand they do interfere, a lot of work has been done on this. FHSS (including 802.11 FH devices) can ruin the performance of a well-tuned DHSS running in the band. IEEE 802.15 TG2 is working on mitigating this in Bluetooth by firmware changes to existing/new kit to reduce likelihood of both transmitting at the same time. Plus in the US, FCC has recently approved changes to band usage that will allow FHSS (incl. Bluetooth) to be more adaptive. Looks like there may be a path to allow OFDM on 2.4GHz soon too (I'd be interested to know if anyone has read anything on OFDM/BT interop? I guess it stands better chance than DSSS since it's multiple carrier - maybe there is still a chance for 802.11g...)
;)
Start here if you want to know more...let's just hope that any modifications to Bluetooth required are done now before many mobile phones are sold. It is next to impossible to get firmware updated on non-tech consumer kit once it's out in the wild (unless there are other, show-stopper, bugs which stop the phone from working correctly - ha, like that's never happened before has it Nokia...
Funny really, I'd have thought Bluetooth would have been better suited to 5GHz. The higher bandwidth available would be useful (both to increase speed and keep bursts short), and the lower range that's possible isn't such a problem for the intended use of Bluetooth.
Of course, 3G licensees and their mobile phone/infrastructure/cash injection suppliers who don't really want free-net hippies eroding the geek local bandwidth market share - after all they have hospitals and cheap booze to pay for - have their own agendas when deciding what tech to push in what band.
Mind you, it's not all bad - personally I'd be quite a bit more interested to see this kind of tech used for the backbone - a neighbourhood is much easier to wire than a long distance link... to get this kind of DX you'll be using fairly tight beams, much less subject to interference, and as long as Bluetooth can be made to tread a little less heavily over local 802.11b base-station installations than it has in the past and share the band fairly then the current generation of cheap(ish) WLAN kit will still be useful and there will be much fun to be had by all...
I hope it does get sorted, things like this could be very cool and while I really don't think d*mned wireless mobile phone headsets are worth the use of free spectrum (*why* can't this kind of thing be run out of *licensed* band if it's run by a commercial service! there should be plenty of that available for lowpower...) I live in hope that everyone'll be able to live together, not stomp all over each other's signal and be happy...
(:
For all but opening new windows, Mozilla is faster than IE on my windows box. Even with the recently added slow-down-turbo-mode-to-accommodate-the-very-few-p eople-who-use-profiles code it's still faster to start.
The Proxomitron is a great adblocker for Windows - lots more modifications possible than with Junkbuster, lots of blocking for certain types of javascript etc., if you want heavy modifications this is definitely the most flexible software. Junkbuster isn't too good at HTTP protocol, you have to fallback to 1.0 which slows things down a lot (see bugzilla).
IE most stable Windows browser... not the way I browse it isn't - current Mozilla builds are usually staying alive for 2-3x as long :)
Paper motherboards, plastic cases? I guess you've only been looking at the low-end NetServers. Some of the rackmounts are lovely, and certainly in the UK they're a lot easier to buy than similarly specced Proliant boxes..
PPP over SSH isn't a very good solution when there's any packet loss (a definite possibility with wireless) since the two layers of TCP interfere. cipe is probably worth a look (NT and Linux).
300 - that was back when modems only transmitted 1 bit per baud, so I guess you can be forgiven... Hey, perhaps that's it - maybe they're talking about 28.8kbaud :-)
Host-header based virtual servers doesn't work with https.