If you are MCI/Verizon then you're not paying MAE/Verizon for a connection. You are paying a mortgage on the building and you are paying for equipment but those costs are more or less the same whether you trade 1 gbps of traffic or 100 gbps of traffic.
In other words, because members of the core move 100% of their traffic via peering they pay no ADDITIVE cost for additional traffic. They still pay a fixed base cost to be in the game but they can increase their traffic without increasing their cost.
Anyway, MAE-East is essentially dead. That peering has all moved to Equinix in Ashburn. But there are other places to peer: Switch & Data PAIX is a good choice. Flat rate access to their switch is reasonable and cross-connects for the folks with which you have enough traffic aren't bad either. At the other end of the spectrum you can peer at a place like NeutralNAP. There are only half a dozen providers there but access to the peering switch is included in the rent.
I'd be really surprised if that box on the wall only cost Verizon $400. I'd bet the total additive cost per customer is in the neighborhood of $2500. That only includes the fiber back to the first multiplexing point. They have trunking costs beyond that.
They have changed hardware, but they didn't go back out to the existing installs. Those they've left in place. They'll probably change one more time, to gpon. I suspect that's half the reason they're not offering business fios in most of the fios-enabled areas yet. They're just waiting for the right equipment.
On the other hand, they can amortize the cost over a number of years. Figure 90% of the installs will continue spending $500/year on the Internet service. In 5 years they've recovered the additive costs and in 10 they're making money. That's if everybody cancels their phone service and switches to VoIP. Perhaps only 40% of the customers will do that. The rest will continue to use Verizon's now less expensive infrastructure. Did I say less expensive? That's right. They no longer have to power your phone for one thing. If you're more than a certain distance from the CO, they no longer have to maintain powered equipment in their CEV's either. (CEV=Controlled Environment Vault). Those suckers are expensive to maintain. Everybody for which they can continue collecting $500/year for phone service represents narrows that break-even to about 3 years.
Now add in the 25% of folks who opt to subscribe to FiOS TV. That's a windfall; they add exactly no cost to provide that. Some of these installs will have covered the additive costs in 2 years and be profitable in 4. They're amortizing on 10-20 years, so that's pretty good news for Verizon.
what after you are a certain size company, you get free connectivity at all the NAPs?
Yes, more or less. Those dozen or so companies are known as the "Internet Core Exchange." They include folks like MCI/Verizon and Abovenet. The process is known as "Reciprocal Peering" and the ICE move 100% of their traffic that way.
FIOS is the internet as it should have been 10 years ago. Every house in America should have had Fiber 10 years ago.
Passive Optical Networking Systems (PONS) weren't around 10 years ago. The technology first saw serious discussion about any form of commercial deployment in 2003.
Without PONS it wasn't cost-feasible. You needed two fibers to every location instead of just one, and the fiber had to go all the way back to powered equipment at the CO instead of being merged with other subscribers on to a small number of fibers that go back to the CO.
DSL and cable ISPs don't support VLANs or IPV6 either
Not so. Do your homework.
There are enough DSL and cable ISPs that I'm sure some of them support IPv6. Very few, but some.
As for VLANs, they work just fine on FiOS as long as you're using one that is OK with a dynamic IP endpoint.
Can you run Skype and Vonage, or are they blocked?
I'm using Vonage, so the answer is yes. Works just fine.
I also ripped out the D-Link router the provide and hooked up my Linux firewall directly to the ONT. This means I get unmolested TCP/IP (except for port 80) all the way to my Linux box.
Its PPoE so I have about a 1480 byte MTU instead of 1500. This has not caused me a problem.
1) they do not remove copper usually. They never will if asked not to.
Perhaps this depends where you are and whether the lines are in the ground or overhead. In Virginia they wanted to remove my overhead copper lines but didn't because I had a T1 at the time. You are correct that they won't insist on removing them if you ask them not to but you have to know to ask.
2) you do not need phone service
However, ordering it without phone service is *really* hard. You will get a serious runaround if you try. You're better off ordering a phone, having it converted to FiOS and then cancelling the phone.
4) Their network absolutely can handle the bandwidth. I can saturate my 15 mbit connection 24/7.
Ditto. Verizon doesn't deserve to win because of how they behaved in the '90s but as much as I hate to say it, they have the best thing going right now.
Equinix is insanely expensive. I considered moving my company's colocation into Equinix's Ashburn VA facility but I ended up choosing more space at two others for half the price. Equinix has a beautiful place but I can't for the life of me figure out who actually needs biometric locks on the cages. That stuff isn't cheap.
You're entirely in a private office park you say. Less than 2 km between buildings, right? Has the telco laid any cable conduit? If so, its now a fixture of the property and belongs to the property owner. This means you can use it. Pick up some spools of direct-burial multimode fiber on ebay at around 20 cents a foot, pull it yourself and pay a fiber expert to come in and attach the connectors.
Even if there is no pre-existing conduit, you can use something like the $250 borit tool to get under the parking lot without disturbing the surface. http://www.borit.com/index.cfm?fuseaction=details& prodID=98 Digging yourself is more expensive than pulling through conduit but it could well be cheaper than a decent laser link and its a whole lot more reliable. And oh yeah -- it'll handle your bandwidth needs for the next 20 years instead of having to be replaced in 3.
Forget 1920. Dvorak hasn't listed a single thing that someone from 1970 wouldn't find bizarre and inexplicable. E-Mail? The Internet? You could fit all the people who understood the concept into a single convention hall.
In Thompson's youth, kids didn't play violent games. They just ran around with toy pistols playing cowboys and indians where they pretended to shoot and kill each other. Well, mostly the pretended to exterminate the Indians because everybody rooted for the cowboys to win.
Of course, they were fully clothed and didn't desecrate any all-american baseball bats along the way, so it was all good clean fun.
No, your IP address may have been included on one of the blacklists, but your server was not "blocked".
That's like a hollywood exec saying, "I didn't blacklist that actress; I just mentioned to everyone I know that she was a communist and let them make up their own minds whether to support communism."
Fine if you're right about her. Libel if you're not.
Spamcop's secret addresses... Not secret enough. I run a huge opt-in political list. Someone entered at least one of spamcop's secret addresses in to my web signup form and damned if I can find it. Short of sending a whole new verify-your-subscription request to everyone, I'm stuck on spamcop's blacklist.
I needed an extra humanities course my last semester in college, so I took a philosophy course. Until then I figured philosophy was pretty useless. I thought that about English courses until the 11th grade too. I was wrong about the English courses. So I figured maybe I was wrong about philosophy too. I had an opportunity to find out and I took it.
The professor described philosphy as the birthplace of science. That was one of only two valuable things I learned in the course.
The course introduced me to a few of the so-called great philosophers. They were, one and all, completely full of shit. Any creative thinker can an interesting discussion about the meaning of life or another such topic. If the discussion moves in a usable direction it can be put to the test with the scientific method. If it doesn't, its pure proselytizing regardless of whether you apply rigorous logic.
Philosphy's utility in Computer Science is even less. We have an added advantage over the pure sciences: the computer tells us pretty quickly when we get things wrong. If we're smart enough to identify the corner cases, it tells us 100% of the time. We don't have to guess or make logical connections. Its all right there in the computer and our ideas either work or they do not.
So the second thing I learned in the philosophy course was this: how to recognize the difference between when scientists or developer knows what they're doing and when they're spouting philosphical bullshit that almost sounds good.
So now I know how to keep the BS out of CS. A useful skill to be sure, but not one that gives me a great deal of respect for philosphy.
You are aware that modern Netware runs under SuSE Linux, right?
I use Samba and I like Samba. If the scope of your task is a small office with 10 users, one fileserver and one printserver then Samba is a fine choice.
I don't think I'd try to set up a clustered san-based fileserver environment using it. I probably could but some of the range of software components I'd have to use are pretty sketchy. When you want "enterprise" filesharing, Netware has it all in one place with refined management GUIs.
I've said it before and I'll say it again: BPL is Internet Fools Gold. The power companies are going to keep pouring money in to this until PONS gets so far ahead of what's theoretically possible with BPL that they finally give up.
Every power line is an antenna, fouling nearby radio with signals placed on it and absorbing signals from nearby radio and noise. Every transformer is a barrier that requires a rugged powered device to bridge the Internet signal for those four housholds. These are fundamental constraints to which no reasonable engineer expects to find a solution.
Why are TV sets, DVD recorders, MP3 players, cell phones, and other software-laden electronic devices reliable and secure but computers are not?
Well, the nice thing about software in rom is that you can't write to it. If you can't inject your own code and unplugging and replugging the device does a full reset back to the factory code then there is a very limited about of damage a hacker can do.
Then too, sets capable of receiving a sophisticated digital signal (HDTV) have only recently come in to wide-spread use. To what extent has anyone even tried to gain control of a TV set's computer by sending malformed data?
Civil liberties in America are no different today than they were pre-9/11.
A couple weeks ago the "miss utility" folks came out and marked where all the fiber optic lines are buried near where I work. The water company wanted to install a new water main without breaking anything.
I want a better Internet link, so I printed a couple satellite pictures off google and went out with a clipboard and pen to mark down where they were.
So here I am, an overweight white guy walking down a public street making notations on a map based on the markings I see on the ground. I was stopped by the police, quizzed about what I was doing and challenged to produce two photo IDs.
One problem is that few developers are motivated to work on bugs
Yeah, this is one for the "no shit sherlock" column. What did you expect to happen when you eliminated the stable/unstable cycle? At a minimum the individual parts of the kernel would achieve stability at different times so that the kernel as a whole was never stable.
This frustrates me immensely at work. I hung on to 2.4 as long as I could. Hardware compatibility pushed me to 2.6 and it just isn't as reliable.
Well, I can't speak for "people like me" but I can tell you that I only challenge borderline messages. Anything that's clearly legit or clearly spam gets delivered or routed to the bit bucket respectively.
I would also suggest that someone who wished you grief could just as easily send forged messages to any of the many thousands of autoresponders out there.
Perhaps you should reserve your ire for the folks who forged your address in the first place... After all that's actually illegal now. You can pursue them in court.
You get those bounces from every single deferred bouncer out there. For example, every qmail installation. You'll only get the challenge from me if the message fell into the grey area between spam and not spam. When SpamAssassin says the message scores a 30, I don't feel a need to double-check. When it says its a 3, I feel the need.
1 spam in 7 generates a challenge because there is only a "moderate" probability that its spam.
5 out of 6 challenges are undeliverable.
So, for every 42 spams I receive, one address accepts a challenge notification.
Now, some of those are actual spammers trying to clean dead addresses from their lists. Others are dead accounts that no one will ever look at. When all is said and done, perhaps 1 in 5 of the challenges accepted by the remote mail system will either get in front of someone or be blocked by their spam filter.
So, for every 200 spams I receive, one individual receives a challenge message from me based on a message where someone forged their address. I include the message headers, so that individual can go beat up on or sue the spammer for forging their address if they want to. I do not include the content, so the spam itself doesn't get propogated. I think that's pretty reasonable.
2. That's pretty rare. I only challenge on messages with a moderate probability of being spam and then only if they're not flagged as being direct from a dynamic IP. Messages flagged as a high probability of being spam go straight to the bit bucket. With the filter in place for more than a year I haven't had a false positive on the high-probability messages yet.
When I get a message with a moderate probability of being spam, my spam blocker sends a message back requesting that the sender confirm the message. Works great. Those few legitimate senders stuck on a problematic server can still get their messages to me and so far no spammer has attempted to bypass it.
The only time it doesn't work is when the sender's spam blocker dumps the confirmation request or when the sender doesn't understand what to do.
Seems to me that any technophobe who wants a FOSS app simplified or wants it to run it on Windows is free to pay the developer of his or her choice to make the port. If the changes are any good they'll probably even be accepted upstream.
Slashdot Mirror
If you are MCI/Verizon then you're not paying MAE/Verizon for a connection. You are paying a mortgage on the building and you are paying for equipment but those costs are more or less the same whether you trade 1 gbps of traffic or 100 gbps of traffic.
In other words, because members of the core move 100% of their traffic via peering they pay no ADDITIVE cost for additional traffic. They still pay a fixed base cost to be in the game but they can increase their traffic without increasing their cost.
Anyway, MAE-East is essentially dead. That peering has all moved to Equinix in Ashburn. But there are other places to peer: Switch & Data PAIX is a good choice. Flat rate access to their switch is reasonable and cross-connects for the folks with which you have enough traffic aren't bad either. At the other end of the spectrum you can peer at a place like NeutralNAP. There are only half a dozen providers there but access to the peering switch is included in the rent.
I'd be really surprised if that box on the wall only cost Verizon $400. I'd bet the total additive cost per customer is in the neighborhood of $2500. That only includes the fiber back to the first multiplexing point. They have trunking costs beyond that.
They have changed hardware, but they didn't go back out to the existing installs. Those they've left in place. They'll probably change one more time, to gpon. I suspect that's half the reason they're not offering business fios in most of the fios-enabled areas yet. They're just waiting for the right equipment.
On the other hand, they can amortize the cost over a number of years. Figure 90% of the installs will continue spending $500/year on the Internet service. In 5 years they've recovered the additive costs and in 10 they're making money. That's if everybody cancels their phone service and switches to VoIP. Perhaps only 40% of the customers will do that. The rest will continue to use Verizon's now less expensive infrastructure. Did I say less expensive? That's right. They no longer have to power your phone for one thing. If you're more than a certain distance from the CO, they no longer have to maintain powered equipment in their CEV's either. (CEV=Controlled Environment Vault). Those suckers are expensive to maintain. Everybody for which they can continue collecting $500/year for phone service represents narrows that break-even to about 3 years.
Now add in the 25% of folks who opt to subscribe to FiOS TV. That's a windfall; they add exactly no cost to provide that. Some of these installs will have covered the additive costs in 2 years and be profitable in 4. They're amortizing on 10-20 years, so that's pretty good news for Verizon.
what after you are a certain size company, you get free connectivity at all the NAPs?
Yes, more or less. Those dozen or so companies are known as the "Internet Core Exchange." They include folks like MCI/Verizon and Abovenet. The process is known as "Reciprocal Peering" and the ICE move 100% of their traffic that way.
FIOS is the internet as it should have been 10 years ago. Every house in America should have had Fiber 10 years ago.
Passive Optical Networking Systems (PONS) weren't around 10 years ago. The technology first saw serious discussion about any form of commercial deployment in 2003.
Without PONS it wasn't cost-feasible. You needed two fibers to every location instead of just one, and the fiber had to go all the way back to powered equipment at the CO instead of being merged with other subscribers on to a small number of fibers that go back to the CO.
DSL and cable ISPs don't support VLANs or IPV6 either
Not so. Do your homework.
There are enough DSL and cable ISPs that I'm sure some of them support IPv6. Very few, but some.
As for VLANs, they work just fine on FiOS as long as you're using one that is OK with a dynamic IP endpoint.
Can you run Skype and Vonage, or are they blocked?
I'm using Vonage, so the answer is yes. Works just fine.
I also ripped out the D-Link router the provide and hooked up my Linux firewall directly to the ONT. This means I get unmolested TCP/IP (except for port 80) all the way to my Linux box.
Its PPoE so I have about a 1480 byte MTU instead of 1500. This has not caused me a problem.
1) they do not remove copper usually. They never will if asked not to.
Perhaps this depends where you are and whether the lines are in the ground or overhead. In Virginia they wanted to remove my overhead copper lines but didn't because I had a T1 at the time. You are correct that they won't insist on removing them if you ask them not to but you have to know to ask.
2) you do not need phone service
However, ordering it without phone service is *really* hard. You will get a serious runaround if you try. You're better off ordering a phone, having it converted to FiOS and then cancelling the phone.
4) Their network absolutely can handle the bandwidth. I can saturate my 15 mbit connection 24/7.
Ditto. Verizon doesn't deserve to win because of how they behaved in the '90s but as much as I hate to say it, they have the best thing going right now.
Equinix is insanely expensive. I considered moving my company's colocation into Equinix's Ashburn VA facility but I ended up choosing more space at two others for half the price. Equinix has a beautiful place but I can't for the life of me figure out who actually needs biometric locks on the cages. That stuff isn't cheap.
Have you considered fiber? Consider it again.
& prodID=98 Digging yourself is more expensive than pulling through conduit but it could well be cheaper than a decent laser link and its a whole lot more reliable. And oh yeah -- it'll handle your bandwidth needs for the next 20 years instead of having to be replaced in 3.
You're entirely in a private office park you say. Less than 2 km between buildings, right? Has the telco laid any cable conduit? If so, its now a fixture of the property and belongs to the property owner. This means you can use it. Pick up some spools of direct-burial multimode fiber on ebay at around 20 cents a foot, pull it yourself and pay a fiber expert to come in and attach the connectors.
Even if there is no pre-existing conduit, you can use something like the $250 borit tool to get under the parking lot without disturbing the surface. http://www.borit.com/index.cfm?fuseaction=details
Forget 1920. Dvorak hasn't listed a single thing that someone from 1970 wouldn't find bizarre and inexplicable. E-Mail? The Internet? You could fit all the people who understood the concept into a single convention hall.
In Thompson's youth, kids didn't play violent games. They just ran around with toy pistols playing cowboys and indians where they pretended to shoot and kill each other. Well, mostly the pretended to exterminate the Indians because everybody rooted for the cowboys to win.
Of course, they were fully clothed and didn't desecrate any all-american baseball bats along the way, so it was all good clean fun.
No, your IP address may have been included on one of the blacklists, but your server was not "blocked".
That's like a hollywood exec saying, "I didn't blacklist that actress; I just mentioned to everyone I know that she was a communist and let them make up their own minds whether to support communism."
Fine if you're right about her. Libel if you're not.
Spamcop's secret addresses... Not secret enough. I run a huge opt-in political list. Someone entered at least one of spamcop's secret addresses in to my web signup form and damned if I can find it. Short of sending a whole new verify-your-subscription request to everyone, I'm stuck on spamcop's blacklist.
I needed an extra humanities course my last semester in college, so I took a philosophy course. Until then I figured philosophy was pretty useless. I thought that about English courses until the 11th grade too. I was wrong about the English courses. So I figured maybe I was wrong about philosophy too. I had an opportunity to find out and I took it.
The professor described philosphy as the birthplace of science. That was one of only two valuable things I learned in the course.
The course introduced me to a few of the so-called great philosophers. They were, one and all, completely full of shit. Any creative thinker can an interesting discussion about the meaning of life or another such topic. If the discussion moves in a usable direction it can be put to the test with the scientific method. If it doesn't, its pure proselytizing regardless of whether you apply rigorous logic.
Philosphy's utility in Computer Science is even less. We have an added advantage over the pure sciences: the computer tells us pretty quickly when we get things wrong. If we're smart enough to identify the corner cases, it tells us 100% of the time. We don't have to guess or make logical connections. Its all right there in the computer and our ideas either work or they do not.
So the second thing I learned in the philosophy course was this: how to recognize the difference between when scientists or developer knows what they're doing and when they're spouting philosphical bullshit that almost sounds good.
So now I know how to keep the BS out of CS. A useful skill to be sure, but not one that gives me a great deal of respect for philosphy.
You are aware that modern Netware runs under SuSE Linux, right?
I use Samba and I like Samba. If the scope of your task is a small office with 10 users, one fileserver and one printserver then Samba is a fine choice.
I don't think I'd try to set up a clustered san-based fileserver environment using it. I probably could but some of the range of software components I'd have to use are pretty sketchy. When you want "enterprise" filesharing, Netware has it all in one place with refined management GUIs.
I've said it before and I'll say it again: BPL is Internet Fools Gold. The power companies are going to keep pouring money in to this until PONS gets so far ahead of what's theoretically possible with BPL that they finally give up.
Every power line is an antenna, fouling nearby radio with signals placed on it and absorbing signals from nearby radio and noise. Every transformer is a barrier that requires a rugged powered device to bridge the Internet signal for those four housholds. These are fundamental constraints to which no reasonable engineer expects to find a solution.
Why are TV sets, DVD recorders, MP3 players, cell phones, and other software-laden electronic devices reliable and secure but computers are not?
Well, the nice thing about software in rom is that you can't write to it. If you can't inject your own code and unplugging and replugging the device does a full reset back to the factory code then there is a very limited about of damage a hacker can do.
Then too, sets capable of receiving a sophisticated digital signal (HDTV) have only recently come in to wide-spread use. To what extent has anyone even tried to gain control of a TV set's computer by sending malformed data?
Civil liberties in America are no different today than they were pre-9/11.
A couple weeks ago the "miss utility" folks came out and marked where all the fiber optic lines are buried near where I work. The water company wanted to install a new water main without breaking anything.
I want a better Internet link, so I printed a couple satellite pictures off google and went out with a clipboard and pen to mark down where they were.
So here I am, an overweight white guy walking down a public street making notations on a map based on the markings I see on the ground. I was stopped by the police, quizzed about what I was doing and challenged to produce two photo IDs.
Pre-9/11 police would not have dared.
perhaps your hardware isn't so compatible with Linux.
My latest problem is the newer Intel Xeon processors with 2 megs of L3 cache. Kernel 2.4 only sees the 512k L2 cache.
I'm pretty comfortable with the assertion that kernel 2.4 has fallen far enough behind in hardware compatibility that a move to 2.6 is necessary.
One problem is that few developers are motivated to work on bugs
Yeah, this is one for the "no shit sherlock" column. What did you expect to happen when you eliminated the stable/unstable cycle? At a minimum the individual parts of the kernel would achieve stability at different times so that the kernel as a whole was never stable.
This frustrates me immensely at work. I hung on to 2.4 as long as I could. Hardware compatibility pushed me to 2.6 and it just isn't as reliable.
Well, I can't speak for "people like me" but I can tell you that I only challenge borderline messages. Anything that's clearly legit or clearly spam gets delivered or routed to the bit bucket respectively.
I would also suggest that someone who wished you grief could just as easily send forged messages to any of the many thousands of autoresponders out there.
Perhaps you should reserve your ire for the folks who forged your address in the first place... After all that's actually illegal now. You can pursue them in court.
You get those bounces from every single deferred bouncer out there. For example, every qmail installation. You'll only get the challenge from me if the message fell into the grey area between spam and not spam. When SpamAssassin says the message scores a 30, I don't feel a need to double-check. When it says its a 3, I feel the need.
Just to clarify:
1 spam in 7 generates a challenge because there is only a "moderate" probability that its spam.
5 out of 6 challenges are undeliverable.
So, for every 42 spams I receive, one address accepts a challenge notification.
Now, some of those are actual spammers trying to clean dead addresses from their lists. Others are dead accounts that no one will ever look at. When all is said and done, perhaps 1 in 5 of the challenges accepted by the remote mail system will either get in front of someone or be blocked by their spam filter.
So, for every 200 spams I receive, one individual receives a challenge message from me based on a message where someone forged their address. I include the message headers, so that individual can go beat up on or sue the spammer for forging their address if they want to. I do not include the content, so the spam itself doesn't get propogated. I think that's pretty reasonable.
1. That's not a false-positive for me.
2. That's pretty rare. I only challenge on messages with a moderate probability of being spam and then only if they're not flagged as being direct from a dynamic IP. Messages flagged as a high probability of being spam go straight to the bit bucket. With the filter in place for more than a year I haven't had a false positive on the high-probability messages yet.
Your whackamole solution doesn't work either. Too many zombies at otherwise legitimate organizations. Would you victimize them even more?
When I get a message with a moderate probability of being spam, my spam blocker sends a message back requesting that the sender confirm the message. Works great. Those few legitimate senders stuck on a problematic server can still get their messages to me and so far no spammer has attempted to bypass it.
The only time it doesn't work is when the sender's spam blocker dumps the confirmation request or when the sender doesn't understand what to do.
Seems to me that any technophobe who wants a FOSS app simplified or wants it to run it on Windows is free to pay the developer of his or her choice to make the port. If the changes are any good they'll probably even be accepted upstream.