That looks like exactly the sort of thing I'm looking for! Now to see if I can get it working...
Coda *is* a blast. My favorite part is the disconnected operation. My least favorite part is the local caching. If you want to copy a 2GB file to your local machine, that means you need 2GB of free space to hold the file *plus* 2GB of free space to hold the locally-cached copy.
I guess for an essentially academic project it's kinda cool, and for other situations -- like as the back-end for a cluster of web servers -- it would work really well, but for me, where I *am* reading/writing big files around a lot, that part of it really sucks.
Mmmm, if you've already Copyrighted and published the source code, and someone tried to come along and patent the idea, you'd have this little thing called "prior art."
of course, the way the patent system works, that wouldn't mean much...
1. Full read/write for every client who has access to the device. Probably not exabytes of data per second going across it, but certainly megabyes per day back and forth.
2. I'd like it to be roughly equivalent to NFS. i.e. far from perfect but more or less effective.
3. I wouldn't be using the space for anything "mission critical." It's more an issue of it seems like a waste to have as much as several gigabytes on my network going to waste because it's all in tiny little pieces and it would be useful for tossing stuff out there to use temporarily or for temp space while I'm ripping one of my CDs or vice-versa temp space for ISO images, etc. Ephemeral data for the most part.
4. Not that I can forsee.
GFS, from the way I understand it, requires that all disk devices need to be shared between multiple servers by physical bus connections and specifically SCSI connections. A lot of the "wasted" space on my network is IDE. Further, it's disk space that might have other assorted crap on it so I'd like to be able to make a subdirectory on that device that I could then somehow export.
That's why Coda has been particularly attractive -- I can just make a directory on the offending hard drive and export it as part of my Coda volume. If it weren't for the local cache "issue" Coda would actually suit all my needs quite well.
Coda is still a server-based network filesystem. It's just one that you can add multiple servers to the same "neighborhood" so that if you have three servers in your Coda neighborhood, each with 1GB of space allocated to Coda filesystem, you could have a single Coda share of 3GB that would be distributed across all machines. Alternately, you could have a volume of 1GB that would be replicated across all three machines and if one went down, the other two would pick up the load from the machine that went down. (Sort of like software-driven, networked RAID 5.)
The "problem" with Coda, at least from my perspective, and please follow the link in the original article for full details, is that any file accessed by a Coda client is first copied to the local workstation *in its entirety* to be worked on. That means if you need to work on a 2GB file on the file server, you'd better have 2GB of space allocated to Coda cache on your workstation. Kinda sucks.
Because the amount of free space on each individual machine is rather small. In conglomeration it would be a useful volume, but individually each one is kind of wussy.
Actually, you'd be surprised at the amount of data I move across my home network. Enough that I've just had to buy a 24-port 10/100 switch to deal with it all effectively...:)
I really wish Coda *would* do the trick, and certainly wouldn't be overkill for me, if only because I also have a laptop that I take with me on consulting jobs and I **love** the ability to have disconnected access to the fileserver.
It's just that the local cache "issue" bites me in the ass with the amount of data I *do* move around my network.
Inter-mezzo has been mentioned several times in response on this article and if I can ever get the website to respond, I'll certainly check it out.
A friend of mine has actually done this. The problem with NBD and md is that it seems after either a certain amount of activity, either over a long period of time or short periods of heavy activity, the NBD on any given machine may decide to head south and go away until the machine gets reset.
While I don't want to use the "Gathered" disk space from all these machines for any sort of "mission critical" purpose, I would like it to at least be fairly stable under average loads.:)
I'm currently using NFS and autmount on all my machines as it is -- just with several shares off of my "real" file server. (Which has something like 35GB of software-RAID on it that gets backed up regularly, is used for "mission-critical" file storage, etc. In other words, it's a "real" file server.)
My problem with using NFS for this is that it's more or less useless if it's broken up into lots of little shares because the individual "pieces" aren't all that big. It's probably several gigabytes alltogether though, which makes it at least potentially useful.
No, I've thought it through quite well. I have several machines on my home network that have very well-defined functions and therefore have fairly well-defined disk usage requirements. The "extra" disk space on the network would not be used for any "mission-critical" application/data storage -- just junk space that's useless as several small partitions that might be somewhat useful as one sort-of large partition. Anything allocated on that space would be entirely ephemeral and I wouldn't care one bit if it vaporized into the bit bucket. I just hate to see as much as several gigabytes going to waste just because it's not all in one place.
I've been an on-and-off shortwaver/world-band guy for a couple of years now, but this, I think, might be the absolutely coolest thing I've ever seen on slashdot. I'll definitely have to uncork my little Realistic world-band receiver and put up the big ol' antenna again after reading this story.
"On the other hand, if the encryption is done at layer 2... They have to decrpyt every single packet you send looking for gold..."
Yeah, except there's a flaw in this logic: if you're encrypting everything at a low layer, the network has no idea where to send the packets because their destination addresses are all encrypted. So, the network hardware needs to be able to read the destination, which means the FBI, CIA, NSA, McDonalds can just check the source/destination on the packets the same way the network does.
To get around anything like this you'd have to A) have a globally-secure Internet that has absolutel yno back-doors for the feds to sneak in (yeah right) or B) a global VPN that connected every network with every other network directly over encrypted links so you'd have to know beforehand where traffic was coming from/going to so you could snoop that specific connection, which of course defeats the whole purpose of a routable network.
I'm avidly waiting ubiquitous high-bandwidth home connections and easy IP telephony. Combine the two with strong encryption and the guv'ment can go sit-n-spin. They can have my strong encryption when they pry it from my cold, dead hands. (Which, unfortunately, will probably be an action put into law sometime by 2001.)
Already we have things like PGPfone, but, well, to be perfectly honest, it really sounds like you're using a cheap microphone hooked up to the Soundblaster on your PC. Anyone know of anything in this realm that actually works well?
I believe I read this in the intro to a new edition of Ender's Game -- I hope I've remembered correctly: Card is a devout Mormon and IIRC has even made missionary work a big part of his personal/religious life. In response to one of the reviewer's comments on Card's tendency towards "faith" in his writing, I have to say, considering his personal/religious background, I think he does a darn good job of not being too preachy or didactic in his writing. Reference Heinlein for someone who does import a lot of his own personal beliefs into his writings. Not that this is necessarily a bad nor a good thing, in either case.
(And this isn't supposed to be a dig against against Mormons; just from what I understand of the religion, missionary work/conversion work is a rather large part of it, and one would sort of expect a devout Mormon author to impart a lot of his belief system in his writing simply as a matter of fact.)
Can we finally put this whole Amiga thing to rest now that it's plainly obvious they're not going to be doing anything anymore?
Why does this remind me of nothing quite so much as some sort of bizarre Monty Python or Marx Brothers skit?
Groucho: So, you're going to build the new Amiga hardware? Chico: A-yes-a, but-a you see, we was going to build the hardware, but now we no building the hardware. Groucho: So you're building the software, right? Chico: Oh, no, you-a see, we-a going to build the soft-a-ware, but now we-a no building the soft-a-ware neither. Groucho: But you're at least designing the chips? Chico: No, we-a thought we-a design the chips, but we then think, no, we no design-a the chips. (in rapid fire) Groucho: OS? Chico: No Groucho: Keyboard? Chico: No Groucho: Mouse? Chico: No Groucho: I've got it! You're going to just build the hardware specifications! Chico: No, we don' going to have-a nothing to do with-a building the boxes and-a we no going to have-a nothing to do with the software that-a running on-a the boxes. Groucho: Well, then what in the world is going to make it an Amiga? Chico: Ah, you-a see, we put-a the Amiga name on it! That-a make it Amiga! Groucho: (pauses) My friend, that is the stupidest idea I have ever heard. And I've heard some pretty stupid ones from you. Chico: Thank-a you very much! We think it-a pretty good too.
Why is this funny and insightful? It reads way too much like the excerpts from "Microserfs" that floated around on- and off-line for a bit before the book came out -- sounded like "Good juicy stuff" about life at Microsoft and then turned out to be complete fiction.
So this article is about a guy who has a $200/hr job yet prefers to make money (at $200/day) with his big eBay addiction, develops a hankerin' to collect "antique" electronic hand calculators, rubs elbows with Tom Arnold and was mere moments away from having the opportunity to doink Tori Spelling while at the same time having a completely dysfunctional relationship with wife and daughter? It just sounds way too "cool" for me to really believe it's at all true; more like a character from some Neil Stephenson short story.
It's maybe slightly insightful, but we've known that people get "Addicted" to certain websites for about as long as, well, around about as long as slashdot has been on the scene.
Apple did go a route like this. Then Steve Jobs came back and took the helm and, since the Newton was Scully's idea and Jobs hated Scully with a passion, Jobs pulled the newly spun-off Newton Inc back under Apple's control and killed it off with extreme prejudice.
I don't think Apple's killing the Newton division had anything to do with managing their budget. After all, they had just spun off the Newton division as a seperate company that had no books/accounting in common with Apple any more! No, I think Jobs is merely a petty little child and killed the Newton because it was the one thing his successor/predecessor created that actually was new, innovative and mildly successful without Jobs having a finger in.
Betcha the NetBSD folks have it running on the PSX2 within a few months of it hitting the US market. (Or maybe within a few months of hitting the Japanese market if they're actually crazier than they usually appear to be.)
Is anyone other than me upset that this box has a big, blue "PS2" label on it? I remember what happened to the last series of electronics devices that had a Big Blue "PS/2" label on them...
Almost certainly this is nothing more than all the vendors raising prices. Prices tend to go up around summertime and stay there through the Christmas buying season. They'll go down again after the new year and after everyone gets done buying the computer stuff they didn't get for xmas.
Wonder if the hardware is capable enough to run the ARM Linux port? The PLEB project seems to be a pretty bare-bones ARM system, so a Nintendo ARM-based system couldn't be much less. Mmmm....
I think probably what Alan meant was that, up until recently with IBM releasing a PPC mobo spec, the only companies to really make a "PPC box for the masses" have been Apple and Be, and of course Apple doesn't really count since they don't want to release hardware specs for their machines, so the PPC boxen have been relegated to a sort of backwater where only freaks like the guys at Be and the LinuxPPC folks like to play in.
There *is* a difference between what you buy in the box with the "Official RedHat Linux" name on it and what you download over the 'net and slap onto a CDROM. It's called _support_. If you expect to be able to get support for your brand new purchase of "RedHat Linux" and then come to find out that you instead get a CD-R with the label "RedHat Linux" written on it with a Sharpie an no "built-in" support options, you are likely to be very upset. (This is not to say that every vendor that slaps a copy of the ftp version of RedHat onto a CDR and writes the name in Sharpie is disreputable, of course. That works just dandy for me, but then, I've never had to call RedHat's support line.)
The *content* is the same, but the *product* is vastly different. I am amazed that nobody here is making the distinction. (Well, ok, not amazed, this is after all a slashdot discussion area, so maybe disappointed is a better word.)
Does anyone really think that Bowie made the decision to distribute via SDMI? Or could it possibly have been his record label that agreed with Microsoft and the Liquid Audio folks to use their formats so show "solidarity" between "The recording industry" and "The computing industry" that they "won't stand for any more of this pirated MP3-based music" or some crap like that.
Bowie, no offense to him, probably doesn't know the difference between SDMI, MP3 and TCP/IP.
Slashdot Mirror
Coda *is* a blast. My favorite part is the disconnected operation. My least favorite part is the local caching. If you want to copy a 2GB file to your local machine, that means you need 2GB of free space to hold the file *plus* 2GB of free space to hold the locally-cached copy.
I guess for an essentially academic project it's kinda cool, and for other situations -- like as the back-end for a cluster of web servers -- it would work really well, but for me, where I *am* reading/writing big files around a lot, that part of it really sucks.
-=-=-=-=-
of course, the way the patent system works, that wouldn't mean much...
-=-=-=-=-
2. I'd like it to be roughly equivalent to NFS. i.e. far from perfect but more or less effective.
3. I wouldn't be using the space for anything "mission critical." It's more an issue of it seems like a waste to have as much as several gigabytes on my network going to waste because it's all in tiny little pieces and it would be useful for tossing stuff out there to use temporarily or for temp space while I'm ripping one of my CDs or vice-versa temp space for ISO images, etc. Ephemeral data for the most part.
4. Not that I can forsee.
GFS, from the way I understand it, requires that all disk devices need to be shared between multiple servers by physical bus connections and specifically SCSI connections. A lot of the "wasted" space on my network is IDE. Further, it's disk space that might have other assorted crap on it so I'd like to be able to make a subdirectory on that device that I could then somehow export.
That's why Coda has been particularly attractive -- I can just make a directory on the offending hard drive and export it as part of my Coda volume. If it weren't for the local cache "issue" Coda would actually suit all my needs quite well.
-=-=-=-=-
The "problem" with Coda, at least from my perspective, and please follow the link in the original article for full details, is that any file accessed by a Coda client is first copied to the local workstation *in its entirety* to be worked on. That means if you need to work on a 2GB file on the file server, you'd better have 2GB of space allocated to Coda cache on your workstation. Kinda sucks.
-=-=-=-=-
-=-=-=-=-
I really wish Coda *would* do the trick, and certainly wouldn't be overkill for me, if only because I also have a laptop that I take with me on consulting jobs and I **love** the ability to have disconnected access to the fileserver.
It's just that the local cache "issue" bites me in the ass with the amount of data I *do* move around my network.
Inter-mezzo has been mentioned several times in response on this article and if I can ever get the website to respond, I'll certainly check it out.
-=-=-=-=-
While I don't want to use the "Gathered" disk space from all these machines for any sort of "mission critical" purpose, I would like it to at least be fairly stable under average loads.
-=-=-=-=-
My problem with using NFS for this is that it's more or less useless if it's broken up into lots of little shares because the individual "pieces" aren't all that big. It's probably several gigabytes alltogether though, which makes it at least potentially useful.
-=-=-=-=-
-=-=-=-=-
And I bet they'll be hitting the market about the same time....
-=-=-=-=-
-=-=-=-=-
Yeah, except there's a flaw in this logic: if you're encrypting everything at a low layer, the network has no idea where to send the packets because their destination addresses are all encrypted. So, the network hardware needs to be able to read the destination, which means the FBI, CIA, NSA, McDonalds can just check the source/destination on the packets the same way the network does.
To get around anything like this you'd have to A) have a globally-secure Internet that has absolutel yno back-doors for the feds to sneak in (yeah right) or B) a global VPN that connected every network with every other network directly over encrypted links so you'd have to know beforehand where traffic was coming from/going to so you could snoop that specific connection, which of course defeats the whole purpose of a routable network.
-=-=-=-=-
Already we have things like PGPfone, but, well, to be perfectly honest, it really sounds like you're using a cheap microphone hooked up to the Soundblaster on your PC. Anyone know of anything in this realm that actually works well?
-=-=-=-=-
(And this isn't supposed to be a dig against against Mormons; just from what I understand of the religion, missionary work/conversion work is a rather large part of it, and one would sort of expect a devout Mormon author to impart a lot of his belief system in his writing simply as a matter of fact.)
-=-=-=-=-
Why does this remind me of nothing quite so much as some sort of bizarre Monty Python or Marx Brothers skit?
Groucho: So, you're going to build the new Amiga hardware?
Chico: A-yes-a, but-a you see, we was going to build the hardware, but now we no building the hardware.
Groucho: So you're building the software, right?
Chico: Oh, no, you-a see, we-a going to build the soft-a-ware, but now we-a no building the soft-a-ware neither.
Groucho: But you're at least designing the chips?
Chico: No, we-a thought we-a design the chips, but we then think, no, we no design-a the chips.
(in rapid fire)
Groucho: OS?
Chico: No
Groucho: Keyboard?
Chico: No
Groucho: Mouse?
Chico: No
Groucho: I've got it! You're going to just build the hardware specifications!
Chico: No, we don' going to have-a nothing to do with-a building the boxes and-a we no going to have-a nothing to do with the software that-a running on-a the boxes.
Groucho: Well, then what in the world is going to make it an Amiga?
Chico: Ah, you-a see, we put-a the Amiga name on it! That-a make it Amiga!
Groucho: (pauses) My friend, that is the stupidest idea I have ever heard. And I've heard some pretty stupid ones from you.
Chico: Thank-a you very much! We think it-a pretty good too.
-=-=-=-=-
imgis.com
doubleclick.net
preferences.com
linkexchange.com
flycast.com
valueclick.com
burstnet.com
nsads.hotwired.com
ads.msn.com
247media.com
imaginemedia.com
focalink.com
gravbots.com
diradserver.developer.com
ngadcenter.net
hitbox.com
ad*.zdnet.com
*.*/event.ng
click.go2net.com
adfu.blockstackers.com
ad.virtualave.com
members.tripod.com/adm/popup/
link4link.com
adlink.de
newads.cmpnet.com
ads.lycos.com
ads.x10.com
adimages.earthweb.com
ngserve.pcworld.com
startpath.com
ads.freshmeat.net
link4ads.com
geocities.com/toto
a32.g.a.yimg.com
gp.deja.com
/*.*/Ads
/*.*/ads
/*.*/adverts
/*.*/advertising
(Sorry about the adfu stuff there guys, but I Really Hate that banner ad crap...)
FWIW - this blockfile gets probably around 99.99% of all the banner ads I've encountered on the 'net.
-=-=-=-=-
So this article is about a guy who has a $200/hr job yet prefers to make money (at $200/day) with his big eBay addiction, develops a hankerin' to collect "antique" electronic hand calculators, rubs elbows with Tom Arnold and was mere moments away from having the opportunity to doink Tori Spelling while at the same time having a completely dysfunctional relationship with wife and daughter? It just sounds way too "cool" for me to really believe it's at all true; more like a character from some Neil Stephenson short story.
It's maybe slightly insightful, but we've known that people get "Addicted" to certain websites for about as long as, well, around about as long as slashdot has been on the scene.
-=-=-=-=-
I don't think Apple's killing the Newton division had anything to do with managing their budget. After all, they had just spun off the Newton division as a seperate company that had no books/accounting in common with Apple any more! No, I think Jobs is merely a petty little child and killed the Newton because it was the one thing his successor/predecessor created that actually was new, innovative and mildly successful without Jobs having a finger in.
-=-=-=-=-
-=-=-=-=-
-=-=-=-=-
-=-=-=-=-
-=-=-=-=-
-=-=-=-=-
The *content* is the same, but the *product* is vastly different. I am amazed that nobody here is making the distinction. (Well, ok, not amazed, this is after all a slashdot discussion area, so maybe disappointed is a better word.)
-=-=-=-=-
Bowie, no offense to him, probably doesn't know the difference between SDMI, MP3 and TCP/IP.
-=-=-=-=-