I think it's interesting how ballenced the numbers are and how consistant they were over time. 24% destination only pages, 24% source only pages, 30% core pages, and 22% disconnected form the rest. Average hops is also within reason, but really is no surprise. The web is a directed graph. This leads to one way only sections. It means to get back to a page one may need to make a long detour, or just not get back. As for the percentages staying the same. This dosen't surprize me. There hasn't been a fundamental revolution in web layout between the two studies.
5 MBytes of pretty random data is easy to come by.
Use tar to archive up your system, pass the output tar file through gzip or compress. Then encrypt it with tripple DES using a randomly generated key and then forget the key. For better randomness, skip a random number of bytes into the output stream before using. If you want even better randomness, XOR it against another data set. While not perfect, it does generate a huge data set that is hard to predict. Anywhere you can throw in a randomness into the generation method helps improve the output data set. One possibility would be to randomly select files to archive. Randomly you could drop blocks or XOR them with them with the next block. When building a chaff data set like this it's important to lose input information and bring in other randomness as it's built.
For a truely random number generation method see LavaRand.
I felt a little left out. I didn't receive any of the ILOVEYOU emails, but then my ISP (visi.com) had a filter setup rather early and was dropping them as they tried to cross it's email servers.
I did get a first generation copy of the "Mothers Day" version. I did the right thing and passed it on to a bunch of the virus app folks within minutes of receiving it. Other than that one I haven't seen any of them.
I'm seeing 45k/sec page load rates on large pages, not bad. My peak DSL download rate is about 68k/sec. What's even nicer is the turn around time is way down. I'm noticing a significant decrease in the time from page request to it's starting to download. I need to check this durring peak time, but it feels much faster than even last night at this time.
If they reset the timeouts properly on the DNS tables they should be only off the air a few minutes for IP# change over. Basically if you know you are going to be changing a name to IP# mapping and you want good continuity you can reset the timeout values for the name low. This means your DNS server will get hammered harder, but when you finally change from one number to the other it will take effect much faster. Depending on site the normal timeouts for name caching are on the order of hours to days. You can change them down to minutes if you wish for a domain or name. It's just your DNS server gets to work harder if you do.
"The prompts are wrong. Do not run emailed executables, especially from people you know and trust."
Here here. I've been just saying to friends not to run emailed programs for ages. With some of the varients of ILOVEYOU that are possible, you can easily trick people into playing that sound bite/vbs worm. I think now we're going to have to add data to the forbidden email attachments. At this point people really need to examine every attachement to see if it is good or bad.
Actually it was worse, they made pipes out of lead, but then there are many places in the US (Like Minneapolis) and other places around the world that still have lots of lead pipes being used for water distribution. Minneapolis is to the point where it is only the pipes from the mains to the houses, but...
On to Blue Tooth, etc.. One thing that worries me is protocals, extensibility, and controls over extensions. It's my feeling that for the first decade or so you will see lots of little incompatibilities between appliance lines. As time wears on the extensions that cause the incompatibilities will slowly become universal or be dropped as dead ends. Blue Tooth isn't going to make all the appliances talk seamlessly at first. Their will be hickups, and your dealing with an environment where the item will not be replaced for years, possibly decades. It will take time to determin how best to comunicate and what information needs to be transfered.
Something needs to be done about expired domain name handling. Something like a grace period needs to be instituted. It would start after the domain is stopped being served by the root name servers. Durring the grace period only the owner could reregister it. This isn't the first time I've heard of a domain expiring and being grabbed by another before the original owner could reregister it. As for how long the grace period should be? I'm leaning twards 1 month + 1 month per year it was held up to a maximum of 6 months or 1 year.
When I first read this article I was thinking along similar lines. How will this change all the calculations that rely on G? I think it's time to do some recrunching of numbers. The area I think would be most interested is astronomy with the orbits of all the Potentially Hazardous Asteroids. My favorit PHA is AN10 which will pass.0026AU from earth or in other words closer than the moon. You have to wait till 2027 for that to happen. Other close pass bys will happen before then. May 7th is when EH26 will visit at only 0.0406AU.
People have already released viruses like this into the wild. One changes "yes" to "no", another removes "not"s from text, another changes spreadsheet values. I'm sure one of them could be altered for the task.
Didn't we just have an article about liable follies in Englend. The US isn't the only one with plain dumb laws. Face it, pretty much every country has some fucked up laws on the books.
"But it could be put to use in the US, where Judge Feu Rosa says he is in discussion with insurance companies to set up a mobile system to resolve disputes over traffic accidents."
Something tells me the ambulence chasers won't let this happen.
No you're not being paranoid, it's being very realistic. It's protecting the rights of the few from the tyrany of the masses, and that is one of the major things the framers of the constitution wanted to do. Remember this nation (United States of America) was formed by people who were oppressed in other nations. Think very carefully about supporting anything limiting a person's rights. It may be your rights that are limited by the next bill.
Every interleving method I've seen implemented gave each bank of ram it's own set of control and data lines. When an access was fired off it was done to both banks. Bank one was used for odd addressed memory words and bank two got even addressed memory words. When the data cam back from the ram it was all loaded into the MB cache. On average (asuming random accesses, the next word of memory is in the cache hald the time. In pratice code accesses are helped the most as one uses long sequential words of memory. Data acheives a better than 1.5x speedup as often you have data locality as in stack frames and records.
I personally want.nerd and.geek..zoo would also be fun. I can see a use for.null to. Somebody would just love to get.31337 or.hack. Another killer one would be.game.
I've been looking at the costs associated with setting up a server farm for a web site. Ouch! Either you go with a 5U cases and get a cheeper servers but then pay tripple the rack costs, or use 1U servers and pay much more for the server, but lower your rack costs. In the long run the lower rack costs add up fast. I'm looking at a configuration of four rack servers, 1 Web, 1 DB, and 2 live backups then a keyboard/flat pannel display tray. All that will fit in a 5U space. In about 1 year I'll have paid for the higher costs of the 1U servers in saved rack costs alown. Using 5U high servers I'd have had to buy 5 slots to fit the same functionality in.
Actually they usually give each cpu card it's own backplane. With a 3 slot PCI backplane you could add in the ethernet and SCSI needed for a server. If you lay the set on the side it will fit in a 1U case. You now have about 4/5th the volume of a 1U case for power supply, fans, floppy, harddisks, etc.
What ever happened to interleaving memory banks for more speed? It does raise the pin counts, but package technologies have been developed that mitigate that. I have an old 486DX2-66 motherboard that does interleving between two banks of ram. Each cache load loaded two memory words into the external cache instead of one. It won't lead to better write performance, but read rates will nearly double (You get something like a 1.8x effective increase).
Re:The problem with Rambus compared to SDRAM...
on
Will Rambus Go Bust?
·
· Score: 3
The problem with Rambus compared to SDRAM... is a higher latency. This means the CPU has to wait longer until the memory is read, but can read memory faster from then on. The consequence of this is that compilers would have to be optimized for that kind of memory access - i. e. accessing a few pages is expensive (and slow) under Rambus, slower than under SDRAM. Accessing many pages is more effective.
And that's a really difficult optimization as you basically have to optimize data accesses as well as code. Data is where it is, locality really can't be improved easily. Systems can be tuned to grab a few pages in a row, but then that may still slow you down if you only used data or code on one of them.
Slashdot Mirror
My first thought was it was opertunistic timing. Let's hit Slashdot while it's running on a new host. Bet we'll find some chinks.
I've often wondered howmany sites aren't linked to by any other site and have never been scanned by a search engine. Chances are we'll never know.
I think it's interesting how ballenced the numbers are and how consistant they were over time. 24% destination only pages, 24% source only pages, 30% core pages, and 22% disconnected form the rest. Average hops is also within reason, but really is no surprise. The web is a directed graph. This leads to one way only sections. It means to get back to a page one may need to make a long detour, or just not get back. As for the percentages staying the same. This dosen't surprize me. There hasn't been a fundamental revolution in web layout between the two studies.
Use tar to archive up your system, pass the output tar file through gzip or compress. Then encrypt it with tripple DES using a randomly generated key and then forget the key. For better randomness, skip a random number of bytes into the output stream before using. If you want even better randomness, XOR it against another data set. While not perfect, it does generate a huge data set that is hard to predict. Anywhere you can throw in a randomness into the generation method helps improve the output data set. One possibility would be to randomly select files to archive. Randomly you could drop blocks or XOR them with them with the next block. When building a chaff data set like this it's important to lose input information and bring in other randomness as it's built.
For a truely random number generation method see LavaRand.
I felt a little left out. I didn't receive any of the ILOVEYOU emails, but then my ISP (visi.com) had a filter setup rather early and was dropping them as they tried to cross it's email servers.
I did get a first generation copy of the "Mothers Day" version. I did the right thing and passed it on to a bunch of the virus app folks within minutes of receiving it. Other than that one I haven't seen any of them.
Charity to donate old server to...
Debian gets my vote...
What do others think?
I'm seeing 45k/sec page load rates on large pages, not bad. My peak DSL download rate is about 68k/sec. What's even nicer is the turn around time is way down. I'm noticing a significant decrease in the time from page request to it's starting to download. I need to check this durring peak time, but it feels much faster than even last night at this time.
If they reset the timeouts properly on the DNS tables they should be only off the air a few minutes for IP# change over. Basically if you know you are going to be changing a name to IP# mapping and you want good continuity you can reset the timeout values for the name low. This means your DNS server will get hammered harder, but when you finally change from one number to the other it will take effect much faster. Depending on site the normal timeouts for name caching are on the order of hours to days. You can change them down to minutes if you wish for a domain or name. It's just your DNS server gets to work harder if you do.
Here here. I've been just saying to friends not to run emailed programs for ages. With some of the varients of ILOVEYOU that are possible, you can easily trick people into playing that sound bite/vbs worm. I think now we're going to have to add data to the forbidden email attachments. At this point people really need to examine every attachement to see if it is good or bad.
Roman Plumers...
Actually it was worse, they made pipes out of lead, but then there are many places in the US (Like Minneapolis) and other places around the world that still have lots of lead pipes being used for water distribution. Minneapolis is to the point where it is only the pipes from the mains to the houses, but...
On to Blue Tooth, etc.. One thing that worries me is protocals, extensibility, and controls over extensions. It's my feeling that for the first decade or so you will see lots of little incompatibilities between appliance lines. As time wears on the extensions that cause the incompatibilities will slowly become universal or be dropped as dead ends. Blue Tooth isn't going to make all the appliances talk seamlessly at first. Their will be hickups, and your dealing with an environment where the item will not be replaced for years, possibly decades. It will take time to determin how best to comunicate and what information needs to be transfered.
Something needs to be done about expired domain name handling. Something like a grace period needs to be instituted. It would start after the domain is stopped being served by the root name servers. Durring the grace period only the owner could reregister it. This isn't the first time I've heard of a domain expiring and being grabbed by another before the original owner could reregister it. As for how long the grace period should be? I'm leaning twards 1 month + 1 month per year it was held up to a maximum of 6 months or 1 year.
When I first read this article I was thinking along similar lines. How will this change all the calculations that rely on G? I think it's time to do some recrunching of numbers. The area I think would be most interested is astronomy with the orbits of all the Potentially Hazardous Asteroids. My favorit PHA is AN10 which will pass .0026AU from earth or in other words closer than the moon. You have to wait till 2027 for that to happen. Other close pass bys will happen before then. May 7th is when EH26 will visit at only 0.0406AU.
However they are having this big debate on what size of coconut to use and if it is better to leave the hairs on or shave them off.
People have already released viruses like this into the wild. One changes "yes" to "no", another removes "not"s from text, another changes spreadsheet values. I'm sure one of them could be altered for the task.
Didn't we just have an article about liable follies in Englend. The US isn't the only one with plain dumb laws. Face it, pretty much every country has some fucked up laws on the books.
Something tells me the ambulence chasers won't let this happen.
No you're not being paranoid, it's being very realistic. It's protecting the rights of the few from the tyrany of the masses, and that is one of the major things the framers of the constitution wanted to do. Remember this nation (United States of America) was formed by people who were oppressed in other nations. Think very carefully about supporting anything limiting a person's rights. It may be your rights that are limited by the next bill.
Every interleving method I've seen implemented gave each bank of ram it's own set of control and data lines. When an access was fired off it was done to both banks. Bank one was used for odd addressed memory words and bank two got even addressed memory words. When the data cam back from the ram it was all loaded into the MB cache. On average (asuming random accesses, the next word of memory is in the cache hald the time. In pratice code accesses are helped the most as one uses long sequential words of memory. Data acheives a better than 1.5x speedup as often you have data locality as in stack frames and records.
I personally want .nerd and .geek. .zoo would also be fun. I can see a use for .null to. Somebody would just love to get .31337 or .hack. Another killer one would be .game.
I've been looking at the costs associated with setting up a server farm for a web site. Ouch! Either you go with a 5U cases and get a cheeper servers but then pay tripple the rack costs, or use 1U servers and pay much more for the server, but lower your rack costs. In the long run the lower rack costs add up fast. I'm looking at a configuration of four rack servers, 1 Web, 1 DB, and 2 live backups then a keyboard/flat pannel display tray. All that will fit in a 5U space. In about 1 year I'll have paid for the higher costs of the 1U servers in saved rack costs alown. Using 5U high servers I'd have had to buy 5 slots to fit the same functionality in.
Actually they usually give each cpu card it's own backplane. With a 3 slot PCI backplane you could add in the ethernet and SCSI needed for a server. If you lay the set on the side it will fit in a 1U case. You now have about 4/5th the volume of a 1U case for power supply, fans, floppy, harddisks, etc.
Quick comment, list the skills you know before jobs, and tell how many years of industry experience you have with each.
Do you have even numbers of memory modules? The old ASUS-SP3G MB that I have requires memory modules in pairs for interleving.
What ever happened to interleaving memory banks for more speed? It does raise the pin counts, but package technologies have been developed that mitigate that. I have an old 486DX2-66 motherboard that does interleving between two banks of ram. Each cache load loaded two memory words into the external cache instead of one. It won't lead to better write performance, but read rates will nearly double (You get something like a 1.8x effective increase).
The consequence of this is that compilers would have to be optimized for that kind of memory access - i. e. accessing a few pages is expensive (and slow) under Rambus, slower than under SDRAM. Accessing many pages is more effective.
And that's a really difficult optimization as you basically have to optimize data accesses as well as code. Data is where it is, locality really can't be improved easily. Systems can be tuned to grab a few pages in a row, but then that may still slow you down if you only used data or code on one of them.