I would suggest investing time looking into something such as Cisco 3750 series switches.
I can't speak for the original poster, but in our case we needed VLAN's to be unique per port. That is, VLAN 100 on port 1 should be switched to VLAN 105 on port 2, and VLAN 100 on port 2 should be switched to VLAN 200 on port 3 and 4, and so on. Trivially easy to do in Linux, not so with a 3750. You can do it with VLAN mapping, but you can also buy quite a server for the price of a 3750 and the Advanced IP Services image. Oh and the 3750 supports only 24 VRF-lites, whereas you can run quite a few more OpenVZ instances with routing on a Linux box.
I am in about the same situation. We ended up going with the RPI-1245 for 12 ports in 1U (by adding a 4-port PCI ethernet card). The solution isn't ideal because only the 4 ports on PCI Express are actually provided with enough bus bandwidth, but it's close enough for us. More ports would be nice, of course. The remote management and the BIOS in general isn't as polished as the HP DL140GL2 we use for everything else.
For our next deployment we will likely go with a 1U switch combined with a 1U HP server, and use VLAN's. The problem is that VLAN's then have to be unique across all ports. If someone decides to make a PC in at most 2U with 24 ethernet ports all on PCI Express, we would certainly go for that. At 4U or bigger you can fit more cards in the machines, but bus bandwidth is a problem if you stick with PCI. PC's with 6 PCI-Express x 4 or 6 PCI-X busses are kind of rare.
I'm not trolling, I'm dead serious- and I'm no windows fanboy either. Sadly, Linux still isn't quite ready for Joe Desktop. It's come a long way, and is certainly getting there, but it's just not ready yet. I'm a Linux user myself, and it is definitely poised to hit the (non-enthusiast) desktop scene in a big way, but not just yet. Be patient.
According to my completely unscientific data (web sites that I run etc.), Linux desktop marketshare is actually falling. The company I work for has just bought a support ticket system where the ticket administration interface only works in Internet Explorer, so I'll need to run Windows in a virtual machine now. In a couple of years people will expect to be able to watch HD disks on their computers, and Linux desktop marketshare will fall even further.
Then again, so far my OS marketshare predictions have been completely utterly wrong. Let us hope this one will be wrong too.
A properly-configured mail server will have a valid PTR record.
Define valid. If you just mean a record that you can resolve to an A record with the same IP address, then yes, your statement is true. Quite a few people do stricter checks than that, and it throws away a lot of legitimate mail.
It's not going to sit on a dynamic IP address, either, since you'd have to change MX records every time the address changes.
Change A records, actually, but the lists of dynamic IP addresses are full of non-dynamic addresses. Anything DHCP-assigned or even somewhat close to an address that is DHCP-assigned counts as dynamic, at least to a certainly popular RBL.
Of the 819 that were blocked by rblsmtpd (that'd be ~78% of inbound mail attempts), 609 were blocked for bad reverse DNS (567) or a known dynamic IP address (42). Of the remainder, 133 were listed at Spamhaus, 74 were listed at SpamCop, and 3 were listed at blackholes.us.
So basically 609 of those could easily have been legitimate messages. Your setup may be good for you, but it would throw away way too much legitimate mail for a business. Oh and the "known dynamic IP address" list contains lots of statically assigned addresses.
As an EE, when TFA uses phrases like "[...] 500 megawatts a year.", it gives me that warm fuzzy feeling that the writer really knows science and engineering. (Sarcasm intended)
He does seem to know what he is talking about. Perhaps you should learn to read before you criticise?
The smaller the distance the heads have to travel, the faster your access time. By only using 63% of a 149 GB drive you cut down on the average distance the heads have to travel, increasing your average access time.
Seek time is by far not linear with seek distance on modern disks. Cutting the seek distance in half doesn't cut the seek time in half.
Fine. Buy the 160GB and make a 100GB partition. Use the 60GB only for archival. Now you've got the same latency as the 100GB and can stop trolling here.
Err, no. At 7200RPM, average rotational latency (half a rotation) is 4.2ms. At 5400RPM, it's 5.6ms. Most of the people being called trolls in the comments to this article aren't actually trolling.
But as they describe it, this HHD doesn't look like a flash device, it looks like a normal hard drive.
As far as I can tell, every place where something like this is implied, there is some get-out clause. I am fairly sure that the drive looks like a flash drive and a hard drive put together. The OS has to actively decide to not put data on the hard drive and instead put them on the flash drive. Windows XP won't notice the flash part of the drive at all, and so won't get any benefit.
Around here you can buy power strips with a special "TV" socket. Plug the TV in the TV socket, and the rest (DVD etc.) in the other sockets. As soon as the power strip detects the TV using less than 20W, it powers off the other sockets. At least that way it's only the TV on stand by.
You can also get power strips with a USB cable. They only supply power when they detect voltage on the USB line -- so turn off your computer, and the peripherals turn off too. Unfortunately there are computers which won't turn off USB power, no matter how much you play with the BIOS settings.
So, assuming the article is true, and discounting the impossible (a combustion process oxidizing into CO2 and water at 6000K), what are the possible input fuels?
The article is true, it just isn't describing combustion. It expects high pressure gas to be injected, and turns mechanical energy into electric energy. Relatively easy to do with high efficiency. Getting the high pressure gas is an exercise for the reader.
Well, my point is that with Skype can call any land-line in any country at local rates, all natively supported by client, protocol, and provider seamlessly. I don't doubt it would be possible to do something involving SIP, but why bother when it's been handed to me on a plate?
What is so hard about typing in a server name, a user name, and a password?
I actually use Skype solely for it's paid-for services (SkypeOut, SMS, etc) as I live abroad, and don't want to pay extortionate rates for international calling. This isn't covered in any SIP clients that I'm aware of
It's certainly not convered in the client, but there are lots and lots of providers, and they'll work with any client. With Skype you're stuck with one provider. Can you even bring your old phone number to Skype?
I have no problem with RTG's. It has been conclusively shown that RTG's are safe even when launch fails. What hasn't been shown is what happens when a RTG impacts Earth at 10km/s or more, as would be the case in a gravity assist gone wrong.
Then theres the protocols; if you need to run several IPSec VPNs through ADSL modems things will get tough. IPSec doesn't just use the normal TCP/UDP protocols oh no, it has *special* protocols...
I must admit I have had no trouble with IPSEC through ADSL modems. IPSEC through NAT used to be a big problem, in particular if you wanted several tunnels through the same NAT device. These days everything supports NAT-T, and that's just UDP on port 4500.
So this guy's solution is to chuck more stuff into the atmosphere to counterbalance what we've already chuucked up there. This is madness.
Better solutions welcome. Stopping the chucking isn't a solution that will happen (except naturally, when the planet becomes uninhabitable for humans. That should lower emissions.)
We need to reduce greenhouse gas emissions and increase carbon sequestration (by planting trees).
You can say we need that as much as you want. It won't happen, unless prices for fossil fuels increase dramatically. Unfortunately coal is plentiful and cheap.
By the way, planting trees is not a solution. If you plant enough trees to mitigate CO2 emissions one year, then the next year you have to plant equally many, and so on, until you run out of land. Then you're screwed.
Bike lights are next to useless at night, useless in a city at night (any secondary reflection of a shop light will get noticed 100 times better) and just plain invisible and a waste in daylight.
I'm not sure what kind of lights you are using, but flashing LED lights are very visible.
Slashdot Mirror
What does this mean? You want traffic switched between the two ports?
I want traffic switched between VLAN 100 on port 1 and VLAN 105 on port 2.
What's connected on the other end, hosts? switches?
Switches, I presume, MPLS routers perhaps. They are at other service providers.
I would suggest investing time looking into something such as Cisco 3750 series switches.
I can't speak for the original poster, but in our case we needed VLAN's to be unique per port. That is, VLAN 100 on port 1 should be switched to VLAN 105 on port 2, and VLAN 100 on port 2 should be switched to VLAN 200 on port 3 and 4, and so on. Trivially easy to do in Linux, not so with a 3750. You can do it with VLAN mapping, but you can also buy quite a server for the price of a 3750 and the Advanced IP Services image. Oh and the 3750 supports only 24 VRF-lites, whereas you can run quite a few more OpenVZ instances with routing on a Linux box.
I am in about the same situation. We ended up going with the RPI-1245 for 12 ports in 1U (by adding a 4-port PCI ethernet card). The solution isn't ideal because only the 4 ports on PCI Express are actually provided with enough bus bandwidth, but it's close enough for us. More ports would be nice, of course. The remote management and the BIOS in general isn't as polished as the HP DL140GL2 we use for everything else.
For our next deployment we will likely go with a 1U switch combined with a 1U HP server, and use VLAN's. The problem is that VLAN's then have to be unique across all ports. If someone decides to make a PC in at most 2U with 24 ethernet ports all on PCI Express, we would certainly go for that. At 4U or bigger you can fit more cards in the machines, but bus bandwidth is a problem if you stick with PCI. PC's with 6 PCI-Express x 4 or 6 PCI-X busses are kind of rare.
I'm not trolling, I'm dead serious- and I'm no windows fanboy either. Sadly, Linux still isn't quite ready for Joe Desktop. It's come a long way, and is certainly getting there, but it's just not ready yet. I'm a Linux user myself, and it is definitely poised to hit the (non-enthusiast) desktop scene in a big way, but not just yet. Be patient.
According to my completely unscientific data (web sites that I run etc.), Linux desktop marketshare is actually falling. The company I work for has just bought a support ticket system where the ticket administration interface only works in Internet Explorer, so I'll need to run Windows in a virtual machine now. In a couple of years people will expect to be able to watch HD disks on their computers, and Linux desktop marketshare will fall even further.
Then again, so far my OS marketshare predictions have been completely utterly wrong. Let us hope this one will be wrong too.
A properly-configured mail server will have a valid PTR record.
Define valid. If you just mean a record that you can resolve to an A record with the same IP address, then yes, your statement is true. Quite a few people do stricter checks than that, and it throws away a lot of legitimate mail.
It's not going to sit on a dynamic IP address, either, since you'd have to change MX records every time the address changes.
Change A records, actually, but the lists of dynamic IP addresses are full of non-dynamic addresses. Anything DHCP-assigned or even somewhat close to an address that is DHCP-assigned counts as dynamic, at least to a certainly popular RBL.
Of the 819 that were blocked by rblsmtpd (that'd be ~78% of inbound mail attempts), 609 were blocked for bad reverse DNS (567) or a known dynamic IP address (42). Of the remainder, 133 were listed at Spamhaus, 74 were listed at SpamCop, and 3 were listed at blackholes.us.
So basically 609 of those could easily have been legitimate messages. Your setup may be good for you, but it would throw away way too much legitimate mail for a business. Oh and the "known dynamic IP address" list contains lots of statically assigned addresses.
Maybe we could just... uh... plant some trees?
What do you plan to do when space for trees runs out?
As an EE, when TFA uses phrases like "[...] 500 megawatts a year.", it gives me that warm fuzzy feeling that the writer really knows science and engineering. (Sarcasm intended)
He does seem to know what he is talking about. Perhaps you should learn to read before you criticise?
The smaller the distance the heads have to travel, the faster your access time. By only using 63% of a 149 GB drive you cut down on the average distance the heads have to travel, increasing your average access time.
Seek time is by far not linear with seek distance on modern disks. Cutting the seek distance in half doesn't cut the seek time in half.
Fine. Buy the 160GB and make a 100GB partition. Use the 60GB only for archival. Now you've got the same latency as the 100GB and can stop trolling here.
Err, no. At 7200RPM, average rotational latency (half a rotation) is 4.2ms. At 5400RPM, it's 5.6ms. Most of the people being called trolls in the comments to this article aren't actually trolling.
But as they describe it, this HHD doesn't look like a flash device, it looks like a normal hard drive.
As far as I can tell, every place where something like this is implied, there is some get-out clause. I am fairly sure that the drive looks like a flash drive and a hard drive put together. The OS has to actively decide to not put data on the hard drive and instead put them on the flash drive. Windows XP won't notice the flash part of the drive at all, and so won't get any benefit.
Actually I'm in Denmark, so the power strips are a bit different from that. It seems like the one you link to works the same way as ours.
Around here you can buy power strips with a special "TV" socket. Plug the TV in the TV socket, and the rest (DVD etc.) in the other sockets. As soon as the power strip detects the TV using less than 20W, it powers off the other sockets. At least that way it's only the TV on stand by.
You can also get power strips with a USB cable. They only supply power when they detect voltage on the USB line -- so turn off your computer, and the peripherals turn off too. Unfortunately there are computers which won't turn off USB power, no matter how much you play with the BIOS settings.
Multi-player games of the split-screen type benefit greatly from high definition.
Having said that, I will still use a NAT-ing router (even if it's IPV6-to-IPV6), so that the Russian mob won't be pounding away on my machine 24x7.
Why is that better than the router simply blocking incoming connections by default?
So, assuming the article is true, and discounting the impossible (a combustion process oxidizing into CO2 and water at 6000K), what are the possible input fuels?
The article is true, it just isn't describing combustion. It expects high pressure gas to be injected, and turns mechanical energy into electric energy. Relatively easy to do with high efficiency. Getting the high pressure gas is an exercise for the reader.
With such a tiny combustion chamber, it's entirely possible that the system can be at 6000C without problem.
Which materials do you propose lining the combustion chamber with?
Well, my point is that with Skype can call any land-line in any country at local rates, all natively supported by client, protocol, and provider seamlessly. I don't doubt it would be possible to do something involving SIP, but why bother when it's been handed to me on a plate?
What is so hard about typing in a server name, a user name, and a password?
I actually use Skype solely for it's paid-for services (SkypeOut, SMS, etc) as I live abroad, and don't want to pay extortionate rates for international calling.
This isn't covered in any SIP clients that I'm aware of
It's certainly not convered in the client, but there are lots and lots of providers, and they'll work with any client. With Skype you're stuck with one provider. Can you even bring your old phone number to Skype?
I have no problem with RTG's. It has been conclusively shown that RTG's are safe even when launch fails. What hasn't been shown is what happens when a RTG impacts Earth at 10km/s or more, as would be the case in a gravity assist gone wrong.
But binary seems most appropriate, self-scaling according to the size of the base unit.
I find e more natural.
Then theres the protocols; if you need to run several IPSec VPNs through ADSL modems things will get tough. IPSec doesn't just use the normal TCP/UDP protocols oh no, it has *special* protocols...
I must admit I have had no trouble with IPSEC through ADSL modems. IPSEC through NAT used to be a big problem, in particular if you wanted several tunnels through the same NAT device. These days everything supports NAT-T, and that's just UDP on port 4500.
So this guy's solution is to chuck more stuff into the atmosphere to counterbalance what we've already chuucked up there. This is madness.
Better solutions welcome. Stopping the chucking isn't a solution that will happen (except naturally, when the planet becomes uninhabitable for humans. That should lower emissions.)
We need to reduce greenhouse gas emissions and increase carbon sequestration (by planting trees).
You can say we need that as much as you want. It won't happen, unless prices for fossil fuels increase dramatically. Unfortunately coal is plentiful and cheap.
By the way, planting trees is not a solution. If you plant enough trees to mitigate CO2 emissions one year, then the next year you have to plant equally many, and so on, until you run out of land. Then you're screwed.
Bike lights are next to useless at night, useless in a city at night (any secondary reflection of a shop light will get noticed 100 times better) and just plain invisible and a waste in daylight.
I'm not sure what kind of lights you are using, but flashing LED lights are very visible.