1gb Internet on average cost ISPs about 10% more than 10mb. I'm not sure it's that much of an assumption that they can turn a profit on $65 for 1gb. A study was done looking at customer data usage and going from 10mb to 1gb only increase their peak usage by 10% on average. I'm sure as new services start cropping up to take advantage of 1gb speeds that customers are starting to get, usage will climb faster, but right now, most everything on the Internet is geared around 10mb or less.
Google has the second most connected network in the world, right behind Level 3 with about 1/4 their size. Google is better connected than most Tier 1s, plus they have a lot of peering with Level 3. I'm sure they have good relations as both Google and Level 3 are on a crusade for net neutrality against most of the other Tier 1s.
I guess at some point, "common sense" is now "cherry picking". It's amazing how long of a life you can live when you "cherry pick" when to cross a street. Why, people normally just go running out into the traffic.
cablecos can put 1/3 of the blame on them for having to go through more node splits faster than planned and upgrading all intermediate hops between the CMTSes and the peering point(s) too.
They're doing it wrong. Do what all the small ISPs are now doing, dedicated 1gb fiber to all customers, all leading back to a consolidation chassis that has more bandwidth than all of the ports running full rate, then just size the uplink port to current traffic trends. No more "splitting nodes", just change out the uplink ports, which are cheap. Last I heard, the going rate for a 100gb line card is down to about $6k.
I link should never go over 50% peak utilization. If it does, you need to upgrade. Unless everyone is using UDP at max line rate, a link shouldn't have congestion.
It is not 100% profit since Comcast still needs to add ports to their routers for Netflix to connect to
Which is covered 100% by the customer. Like I said, What Netflix pays is just frosting on the cake.
Netflix discovered that direct peering might actually not be as bad
Paying $0.5/mbit to peer with someone is outrageously expensive. For $0.50, you can get transit from LA to NY, not just down the street like with peering.
Doesn't matter. Even fibers will be saturated one day and then the whole process starts again.
When my 1pb/s fiber connection can no longer go any faster, I guess I'll just be stuck with 1pb/s. Ohh, why yes, didn't you hear? They can do about 900tb/s over a single fiber now. That's 9x peak Internet bandwidth over a single fiber. Gotta love the progress of technology.
It's wasteful of instantaneous bandwidth in a network that currently happens to be saturated.
I've been telling people this for years, caching files in memory is wasteful! Just read it from disk. Think of the savings! Let the HD parks itself to save power? Hell no! We don't want to be wasteful! Make those motors spin 24/7!
Today I can't even call people in the next county without dialing an area code and incurring LD charges!
My uncle is one of the few people who I know that still owns a land line. He pays $10/month for unlimited nationwide with callerid. I didn't know there was still a notion of "long distance" for anything inside the 48 states.
You only need QoS if you have congestion. Get rid of the congestion and you don't need to reserve anything. VoIP is such low bandwidth that it should not matter. The only issue we should address is how to handle DoS attacks.
It's been a while, but I think the break even point for computational costs between RSA and ECC is something around 1500bit RSA and 200bit ECC, but the 200bit ECC is much stronger. ECC's computation grows closer to linear while RSA is exponential. With 2048bit RSA being standard now, ECC is now cheaper, minus hardware acceleration reasons.
It's not "free as in freedom", it removes freedoms, like the freedom to keep your changes to yourself. You take away one person's freedom and give it to another.
I can't agree with forcing people to be altruistic. Now if the argument is that GPL is a necessary evil that contradicts its own agenda, I can appreciate that. What I can't stand is flagrant BS that RMS claims to be "free" software. There is nothing inherently wrong with using GPL, that's up to the person doing the work, but the argument made about being "truly" free is utter crap. It's "mostly" free. BSD even has restrictions, but those less about removing freedoms from the end user and more about getting credit where credit is due.
His definition of "free software" tells users how they can and cannot use the software. His ideal "free" software would not be free except by his made up definition.
$0.3/mbit for peering is EXPENSIVE. Since all of the equipment costs are already covered by their residential customer, that's a 100% net profit. Kind of like if Microsoft started subsidizing their xbox games to price the playstation out of the market. That's called monopolistic power for a reason.
Google Fiber is a separate self sustaining entity... well... should be. We'll see in a few years. The main point is Google Fiber isn't meant to cost Google anything, it is meant to make money directly from its own value, not by added value to the main company.
IPv6 is faster, but with my limit background but decent understanding, am going to say it's only slightly faster in the way of computation, but could have minor benefits now and biggest benefits in the future because of much smaller routing tables. I wouldn't say "20%" faster, because most hardware does IPv4 at line rate, so you can't get any faster, but IPv6 is technically simpler, so it should have a minor benefit.
My guess is the only major benefit is smaller routing tables, so unless that's a limit factor, there should be almost no difference. Mind oyu, fragmentation is on the rise, so the routing tables are growing fast for IPv4.
A lot of companies build for peak loads. Level 3 does because they sell all of their bandwidth as congestion free in their network. There is a difference between theoretical peak and practical peak. theoretical peak is like flipping a coin 100 times and getting heads, while practical peak is based on just staying ahead of current usage.
Actually, Level 3 claims to maintain 2x peak. If the peak is 1gb, they will make sure they have 2gb.
As for last mile ISPs, terabit backplanes are the norm with any fiber purchased in the past 5 years. We now have access to multi-terabit consolidators with 100gb uplinks that will soon be 1tb uplinks when the standard comes out in 2015. Don't forget about the petabit routers they now have.
With residential affordable, one could build a network to handle all of New York city with 1gb connections to every house/apartment, and allow full non-blocking connections at full-duplex 1gb speeds in a full-mesh design, including routing. Complete overkill, but it can be done.
Level 3 does provide the highest quality service for a competitive price, but it doesn't matter because most ISPs have a monopoly. The big player ISPs have no interest in quality, which means Level 3 needs to create motivation for peering, which means lining the golden pockets of the incumbent ISPs even more than they already are.
If changing your ISP involved nothing more than making a 5 minute phone call and getting your VLAN changed from Comcast to someone else, then there would be a reason to keep the customers happy. As it stands right now, getting top notch speed to a colo'd speed test server is about the only amount of quality one can expect. Actually getting out to the Internet is an exercise of patience.
Level 3 is the single largest Tier 1 in the world, several factors larger than the 2nd largest, which is several factors larger than the 3rd largest. Level 3 is crazy big, but owning a multi-billion dollar backbone with razor thin low single digit margins means nothing compared to the trillion dollar last mile owned by incumbent ISPs who get large double digit margins on magnitudes larger revenues.
Except that Comcast has no monopoly -- government allowed, government sponsored, or otherwise
They did in the very recent past. I'm not sure if they still have this, but in many areas, they had franchise agreements that allowed only one cable or telcom providers in an area for upwards of 10 years. Even without franchise agreements, around here, that have city ordinances that were voted on by the people to only allow one of each type of provider because people hate getting their lawns torn up every time a new provider wants to come in.
The problem with being a fixed line ISP, is you also need to get property rights for both public and private property. It's not like a normal business that can setup shop anywhere and just have store front or install a few wifi towers.
According to some other sites, they said 8.0 users will still get updates, but if you have 8.1, you will be required to have 8.1 update 1 to continue getting updates.
Let me get this strait, Level 3 sells bandwidth of the highest quality to a company, routes it around the world with nearly no congestion, then offers to peer with an ISP for free, meaning that ISP doesn't need to route the data around the world themselves, the ISP refuses because they think the data should be not only handed to them on a silver platter, but also get paid; and you think Level 3 has a "horrible" business model?
Slashdot Mirror
1gb Internet on average cost ISPs about 10% more than 10mb. I'm not sure it's that much of an assumption that they can turn a profit on $65 for 1gb. A study was done looking at customer data usage and going from 10mb to 1gb only increase their peak usage by 10% on average. I'm sure as new services start cropping up to take advantage of 1gb speeds that customers are starting to get, usage will climb faster, but right now, most everything on the Internet is geared around 10mb or less.
Google has the second most connected network in the world, right behind Level 3 with about 1/4 their size. Google is better connected than most Tier 1s, plus they have a lot of peering with Level 3. I'm sure they have good relations as both Google and Level 3 are on a crusade for net neutrality against most of the other Tier 1s.
I guess at some point, "common sense" is now "cherry picking". It's amazing how long of a life you can live when you "cherry pick" when to cross a street. Why, people normally just go running out into the traffic.
cablecos can put 1/3 of the blame on them for having to go through more node splits faster than planned and upgrading all intermediate hops between the CMTSes and the peering point(s) too.
They're doing it wrong. Do what all the small ISPs are now doing, dedicated 1gb fiber to all customers, all leading back to a consolidation chassis that has more bandwidth than all of the ports running full rate, then just size the uplink port to current traffic trends. No more "splitting nodes", just change out the uplink ports, which are cheap. Last I heard, the going rate for a 100gb line card is down to about $6k.
I link should never go over 50% peak utilization. If it does, you need to upgrade. Unless everyone is using UDP at max line rate, a link shouldn't have congestion.
Voting doesn't work because the average person is an idiot or too lazy to care, which breeds idiocy.
It is not 100% profit since Comcast still needs to add ports to their routers for Netflix to connect to
Which is covered 100% by the customer. Like I said, What Netflix pays is just frosting on the cake.
Netflix discovered that direct peering might actually not be as bad
Paying $0.5/mbit to peer with someone is outrageously expensive. For $0.50, you can get transit from LA to NY, not just down the street like with peering.
Doesn't matter. Even fibers will be saturated one day and then the whole process starts again.
When my 1pb/s fiber connection can no longer go any faster, I guess I'll just be stuck with 1pb/s. Ohh, why yes, didn't you hear? They can do about 900tb/s over a single fiber now. That's 9x peak Internet bandwidth over a single fiber. Gotta love the progress of technology.
It's wasteful of instantaneous bandwidth in a network that currently happens to be saturated.
I've been telling people this for years, caching files in memory is wasteful! Just read it from disk. Think of the savings! Let the HD parks itself to save power? Hell no! We don't want to be wasteful! Make those motors spin 24/7!
/sarc
Today I can't even call people in the next county without dialing an area code and incurring LD charges!
My uncle is one of the few people who I know that still owns a land line. He pays $10/month for unlimited nationwide with callerid. I didn't know there was still a notion of "long distance" for anything inside the 48 states.
You only need QoS if you have congestion. Get rid of the congestion and you don't need to reserve anything. VoIP is such low bandwidth that it should not matter. The only issue we should address is how to handle DoS attacks.
Our Common Carrier telephone system, at least until the breakup, was the envy of the world. Rates were reasonable
Yes prices plummeted and quality increased after the breakup. Best not to read history with rose glasses.
It's been a while, but I think the break even point for computational costs between RSA and ECC is something around 1500bit RSA and 200bit ECC, but the 200bit ECC is much stronger. ECC's computation grows closer to linear while RSA is exponential. With 2048bit RSA being standard now, ECC is now cheaper, minus hardware acceleration reasons.
You think that's bad, we still have a few Win98+FF1 agent strings coming through.
It's not "free as in freedom", it removes freedoms, like the freedom to keep your changes to yourself. You take away one person's freedom and give it to another.
I can't agree with forcing people to be altruistic. Now if the argument is that GPL is a necessary evil that contradicts its own agenda, I can appreciate that. What I can't stand is flagrant BS that RMS claims to be "free" software. There is nothing inherently wrong with using GPL, that's up to the person doing the work, but the argument made about being "truly" free is utter crap. It's "mostly" free. BSD even has restrictions, but those less about removing freedoms from the end user and more about getting credit where credit is due.
His definition of "free software" tells users how they can and cannot use the software. His ideal "free" software would not be free except by his made up definition.
http://he.net/ Get BGP+IPv6+IPv4 for $0.45/Mbps!
$0.3/mbit for peering is EXPENSIVE. Since all of the equipment costs are already covered by their residential customer, that's a 100% net profit. Kind of like if Microsoft started subsidizing their xbox games to price the playstation out of the market. That's called monopolistic power for a reason.
Google Fiber is a separate self sustaining entity... well... should be. We'll see in a few years. The main point is Google Fiber isn't meant to cost Google anything, it is meant to make money directly from its own value, not by added value to the main company.
IPv6 is faster, but with my limit background but decent understanding, am going to say it's only slightly faster in the way of computation, but could have minor benefits now and biggest benefits in the future because of much smaller routing tables. I wouldn't say "20%" faster, because most hardware does IPv4 at line rate, so you can't get any faster, but IPv6 is technically simpler, so it should have a minor benefit.
My guess is the only major benefit is smaller routing tables, so unless that's a limit factor, there should be almost no difference. Mind oyu, fragmentation is on the rise, so the routing tables are growing fast for IPv4.
A lot of companies build for peak loads. Level 3 does because they sell all of their bandwidth as congestion free in their network. There is a difference between theoretical peak and practical peak. theoretical peak is like flipping a coin 100 times and getting heads, while practical peak is based on just staying ahead of current usage.
Actually, Level 3 claims to maintain 2x peak. If the peak is 1gb, they will make sure they have 2gb.
As for last mile ISPs, terabit backplanes are the norm with any fiber purchased in the past 5 years. We now have access to multi-terabit consolidators with 100gb uplinks that will soon be 1tb uplinks when the standard comes out in 2015. Don't forget about the petabit routers they now have.
With residential affordable, one could build a network to handle all of New York city with 1gb connections to every house/apartment, and allow full non-blocking connections at full-duplex 1gb speeds in a full-mesh design, including routing. Complete overkill, but it can be done.
Level 3 does provide the highest quality service for a competitive price, but it doesn't matter because most ISPs have a monopoly. The big player ISPs have no interest in quality, which means Level 3 needs to create motivation for peering, which means lining the golden pockets of the incumbent ISPs even more than they already are.
If changing your ISP involved nothing more than making a 5 minute phone call and getting your VLAN changed from Comcast to someone else, then there would be a reason to keep the customers happy. As it stands right now, getting top notch speed to a colo'd speed test server is about the only amount of quality one can expect. Actually getting out to the Internet is an exercise of patience.
You peer with Tier 2 ISPs like Level3
Level 3 is the single largest Tier 1 in the world, several factors larger than the 2nd largest, which is several factors larger than the 3rd largest. Level 3 is crazy big, but owning a multi-billion dollar backbone with razor thin low single digit margins means nothing compared to the trillion dollar last mile owned by incumbent ISPs who get large double digit margins on magnitudes larger revenues.
Except that Comcast has no monopoly -- government allowed, government sponsored, or otherwise
They did in the very recent past. I'm not sure if they still have this, but in many areas, they had franchise agreements that allowed only one cable or telcom providers in an area for upwards of 10 years. Even without franchise agreements, around here, that have city ordinances that were voted on by the people to only allow one of each type of provider because people hate getting their lawns torn up every time a new provider wants to come in.
The problem with being a fixed line ISP, is you also need to get property rights for both public and private property. It's not like a normal business that can setup shop anywhere and just have store front or install a few wifi towers.
According to some other sites, they said 8.0 users will still get updates, but if you have 8.1, you will be required to have 8.1 update 1 to continue getting updates.
Let me get this strait, Level 3 sells bandwidth of the highest quality to a company, routes it around the world with nearly no congestion, then offers to peer with an ISP for free, meaning that ISP doesn't need to route the data around the world themselves, the ISP refuses because they think the data should be not only handed to them on a silver platter, but also get paid; and you think Level 3 has a "horrible" business model?