Re: Online games stealing your bandwidth
There's some standardization work being done in the IETF ALTO working group. They're looking at ways ISPs can inform P2P clints about which peers are "better", I.e., topologically nearby. http://tools.ietf.org/wg/alto/ I'm less familiar with DECADE, but I believe they're working on more directly cache-related stuff. http://tools.ietf.org/wg/decade/ On Sep 25, 2010 4:44 PM, "Matthew Walster" <matthew@walster.org> wrote: On 25 September 2010 21:16, Rodrick Brown <rodrick.brown@gmail.com> wrote:
I think most people are... <snip>
I once read an article talking about making BitTorrent scalable by using anycasted caching services at the ISP's closest POP to the end user. Given sufficient traffic on a specified torrent, the caching device would build up the file, then distribute that direct to the subscriber in the form of an additional (preferred) peer. Similar to a CDN or Usenet, but where it was cached rather than deliberately pushed out from a locus. Was anything ever standardised in that field? I imagine with much of P2P traffic being (how shall I put this...) less than legal, it's of questionable legality and the ISPs would not want to be held liable for the content cached there? M
Can someone name an ISP that encourages P2P traffic?? ;) Sent from a mobile phone with a small keyboard, please excuse my mistakes. On Sep 27, 2010, at 4:32 PM, "Richard Barnes" <richard.barnes@gmail.com> wrote:
There's some standardization work being done in the IETF ALTO working group. They're looking at ways ISPs can inform P2P clints about which peers are "better", I.e., topologically nearby. http://tools.ietf.org/wg/alto/
I'm less familiar with DECADE, but I believe they're working on more directly cache-related stuff. http://tools.ietf.org/wg/decade/
On Sep 25, 2010 4:44 PM, "Matthew Walster" <matthew@walster.org> wrote:
On 25 September 2010 21:16, Rodrick Brown <rodrick.brown@gmail.com> wrote:
I think most people are... <snip>
I once read an article talking about making BitTorrent scalable by using anycasted caching services at the ISP's closest POP to the end user. Given sufficient traffic on a specified torrent, the caching device would build up the file, then distribute that direct to the subscriber in the form of an additional (preferred) peer. Similar to a CDN or Usenet, but where it was cached rather than deliberately pushed out from a locus.
Was anything ever standardised in that field? I imagine with much of P2P traffic being (how shall I put this...) less than legal, it's of questionable legality and the ISPs would not want to be held liable for the content cached there?
M
I thought the issue was more about ISPs encouraging *responsible* P2P. On Mon, Sep 27, 2010 at 8:35 PM, Warren Bailey <wbailey@gci.com> wrote:
Can someone name an ISP that encourages P2P traffic?? ;)
Sent from a mobile phone with a small keyboard, please excuse my mistakes.
On Sep 27, 2010, at 4:32 PM, "Richard Barnes" <richard.barnes@gmail.com> wrote:
There's some standardization work being done in the IETF ALTO working group. They're looking at ways ISPs can inform P2P clints about which peers are "better", I.e., topologically nearby. http://tools.ietf.org/wg/alto/
I'm less familiar with DECADE, but I believe they're working on more directly cache-related stuff. http://tools.ietf.org/wg/decade/
On Sep 25, 2010 4:44 PM, "Matthew Walster" <matthew@walster.org> wrote:
On 25 September 2010 21:16, Rodrick Brown <rodrick.brown@gmail.com> wrote:
I think most people are... <snip>
I once read an article talking about making BitTorrent scalable by using anycasted caching services at the ISP's closest POP to the end user. Given sufficient traffic on a specified torrent, the caching device would build up the file, then distribute that direct to the subscriber in the form of an additional (preferred) peer. Similar to a CDN or Usenet, but where it was cached rather than deliberately pushed out from a locus.
Was anything ever standardised in that field? I imagine with much of P2P traffic being (how shall I put this...) less than legal, it's of questionable legality and the ISPs would not want to be held liable for the content cached there?
M
On 9/27/2010 7:35 PM, Warren Bailey wrote:
Can someone name an ISP that encourages P2P traffic?? ;)
A proper ISP doesn't encourage any type of traffic. We're indifferent. Of course, we'll be happy to mention the benefits and draw backs of using various protocols on the Internet. Demand wise, video streaming to point and click boxes will load the network far more than p2p ever has; granted, in the opposite direction of the normal p2p complaint. My, and my company's, biggest complaint is the lack of improvement on these protocols to play more friendly with customer's other traffic. It is not so much the effects of it on my network, as much as how it effects my customer's unshared link. The "give me everything" tactic, especially on outbound traffic, saturates the link, which in turn lowers the customer's other traffic. Am I the only one who likes to stream video while running bittorrent, surfing the web, checking my email, and playing some online game all at the same time? I'm not going to rag on bittorrent, though. I do have adjustments in my clients to cap the upstream/downstream to allow my other traffic through. Many clients and protocols don't have this ability, though. Some purposefully hide themselves and what they are doing. The only indication is the fact that the "Internet is slow." The people who make this software should sit in a call center troubleshooting why "The Internet is slow!" when various software products are bandwidth hogs (and sometimes are hidden from the customer completely). We, of course, detect the link saturation, but there is no indicator for us to help the customer figure out what they need to disable. Jack
BitTorrent have been active contributors to the IETF LEDBAT working group, which is looking at transport protocols that back off much more aggressively than TCP, with exactly the idea of making P2P have a lower impact on other things at the customer edge. <http://tools.ietf.org/wg/ledbat/> On Tue, Sep 28, 2010 at 9:58 AM, Jack Bates <jbates@brightok.net> wrote:
On 9/27/2010 7:35 PM, Warren Bailey wrote:
Can someone name an ISP that encourages P2P traffic?? ;)
A proper ISP doesn't encourage any type of traffic. We're indifferent. Of course, we'll be happy to mention the benefits and draw backs of using various protocols on the Internet. Demand wise, video streaming to point and click boxes will load the network far more than p2p ever has; granted, in the opposite direction of the normal p2p complaint.
My, and my company's, biggest complaint is the lack of improvement on these protocols to play more friendly with customer's other traffic. It is not so much the effects of it on my network, as much as how it effects my customer's unshared link. The "give me everything" tactic, especially on outbound traffic, saturates the link, which in turn lowers the customer's other traffic. Am I the only one who likes to stream video while running bittorrent, surfing the web, checking my email, and playing some online game all at the same time?
I'm not going to rag on bittorrent, though. I do have adjustments in my clients to cap the upstream/downstream to allow my other traffic through. Many clients and protocols don't have this ability, though. Some purposefully hide themselves and what they are doing. The only indication is the fact that the "Internet is slow." The people who make this software should sit in a call center troubleshooting why "The Internet is slow!" when various software products are bandwidth hogs (and sometimes are hidden from the customer completely). We, of course, detect the link saturation, but there is no indicator for us to help the customer figure out what they need to disable.
Jack
Jack, Forgive me if I'm mistaken, but looking at your website - do you only offer dial up services? This could be the background for a statement like "a proper ISP doesn't encourage any type of traffic." We have a couple of OC-192 running to Seattle, so certain "types" of traffic can make a good day turn very badly without some traffic "management". -----Original Message----- From: Jack Bates [mailto:jbates@brightok.net] Sent: Tuesday, September 28, 2010 5:58 AM To: Warren Bailey Cc: Richard Barnes; NANOG Subject: Re: Online games stealing your bandwidth On 9/27/2010 7:35 PM, Warren Bailey wrote:
Can someone name an ISP that encourages P2P traffic?? ;)
A proper ISP doesn't encourage any type of traffic. We're indifferent. Of course, we'll be happy to mention the benefits and draw backs of using various protocols on the Internet. Demand wise, video streaming to point and click boxes will load the network far more than p2p ever has; granted, in the opposite direction of the normal p2p complaint. My, and my company's, biggest complaint is the lack of improvement on these protocols to play more friendly with customer's other traffic. It is not so much the effects of it on my network, as much as how it effects my customer's unshared link. The "give me everything" tactic, especially on outbound traffic, saturates the link, which in turn lowers the customer's other traffic. Am I the only one who likes to stream video while running bittorrent, surfing the web, checking my email, and playing some online game all at the same time? I'm not going to rag on bittorrent, though. I do have adjustments in my clients to cap the upstream/downstream to allow my other traffic through. Many clients and protocols don't have this ability, though. Some purposefully hide themselves and what they are doing. The only indication is the fact that the "Internet is slow." The people who make this software should sit in a call center troubleshooting why "The Internet is slow!" when various software products are bandwidth hogs (and sometimes are hidden from the customer completely). We, of course, detect the link saturation, but there is no indicator for us to help the customer figure out what they need to disable. Jack
On 9/28/2010 1:00 PM, Warren Bailey wrote:
Jack,
Forgive me if I'm mistaken, but looking at your website - do you only offer dial up services? This could be the background for a statement like "a proper ISP doesn't encourage any type of traffic." We have a couple of OC-192 running to Seattle, so certain "types" of traffic can make a good day turn very badly without some traffic "management".
BrightNet itself has ILEC's as customers. We're a turnkey glue for ILECs nearby. Among other things, I provide engineering support and advise for each ILEC. Each has their own levels of service, management, and technologies deployed including wireless, cellular, DSL, FTTH, and cable. I'm currently running around 1.2gbit between us and 4 NSP transits with 3gbit available. Some of the ILECs have additional load shifting with other transits. I estimate the need to go 10Gig ring or split transit in less than 5 years at current growth rates, and the largest problem we've run into is getting infrastructure to handle gig-e speeds out of rural ILECs for the 100+ mile longhauls. I've had issues with gig-e connectivity just getting out of OKC to enough NSP transits and it will become more difficult/expensive when we do hit 10G. As it currently stands, we usually have no problems with event spikes, though we sometimes have to tweek the traffic paths depending on how the NSPs do. The largest issues have always been the last mile. As we resolve last mile costs (which dropping 100% fiber in a rural area today doesn't have the safety nets and guarantees that were provided when copper was dropped in), we'll then have to tackle the longhaul connectivity issues, but hopefully the cost to handle that will drop as well. Jack
Jack, Apologies, I did not realize that you guys were doing so much. Please don't take my last email as anything which was intended to question or insult you guys. Up here (Alaska) we have about 100,000 cable subscribers along with mixed Fiber/DSL/POTS access and nearly 50,000 cellular customers with high speed data around our Metro network. I am an RF Engineer, however the network I run is IP based (satellite) and I run in the neighborhood of 250mbps forward and 30mbps return to most of the State of Alaska. I find that anywhere from 40-65% of our total traffic is "questionable", which is why I was asking about an ISP who liked their users downloading torrents. It is very difficult to gauge a users behavior if they are on an "all out" downloading binge over a weekend. Normally, a user logs in and does what they need to in a relatively short amount of time (hours). In the case of most providers, we oversubscribe our resources and have found this model is beginning to not apply to user behavior changes. Long gone are the days of the user turning off their computers, and our customer base (rural Alaska) have few things to do besides use the internet. This has made a "perfect storm" of sorts, as we are now seeing most of our users utilizing 70%+ of their allocated (purchased) bandwidth 24 hours a day. The vast majority of the night use is gaming, and bit torrent. It makes things much more complicated when trying to give an experience to people.. //warren -----Original Message----- From: Jack Bates [mailto:jbates@brightok.net] Sent: Tuesday, September 28, 2010 10:26 AM To: Warren Bailey Cc: Richard Barnes; NANOG Subject: Re: Online games stealing your bandwidth On 9/28/2010 1:00 PM, Warren Bailey wrote:
Jack,
Forgive me if I'm mistaken, but looking at your website - do you only offer dial up services? This could be the background for a statement like "a proper ISP doesn't encourage any type of traffic." We have a couple of OC-192 running to Seattle, so certain "types" of traffic can make a good day turn very badly without some traffic "management".
BrightNet itself has ILEC's as customers. We're a turnkey glue for ILECs nearby. Among other things, I provide engineering support and advise for each ILEC. Each has their own levels of service, management, and technologies deployed including wireless, cellular, DSL, FTTH, and cable. I'm currently running around 1.2gbit between us and 4 NSP transits with 3gbit available. Some of the ILECs have additional load shifting with other transits. I estimate the need to go 10Gig ring or split transit in less than 5 years at current growth rates, and the largest problem we've run into is getting infrastructure to handle gig-e speeds out of rural ILECs for the 100+ mile longhauls. I've had issues with gig-e connectivity just getting out of OKC to enough NSP transits and it will become more difficult/expensive when we do hit 10G. As it currently stands, we usually have no problems with event spikes, though we sometimes have to tweek the traffic paths depending on how the NSPs do. The largest issues have always been the last mile. As we resolve last mile costs (which dropping 100% fiber in a rural area today doesn't have the safety nets and guarantees that were provided when copper was dropped in), we'll then have to tackle the longhaul connectivity issues, but hopefully the cost to handle that will drop as well. Jack
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 9/28/10 3:01 PM, Warren Bailey wrote:
Jack,
Apologies, I did not realize that you guys were doing so much. Please don't take my last email as anything which was intended to question or insult you guys. Up here (Alaska) we have about 100,000 cable subscribers along with mixed Fiber/DSL/POTS access and nearly 50,000 cellular customers with high speed data around our Metro network. I am an RF Engineer, however the network I run is IP based (satellite) and I run in the neighborhood of 250mbps forward and 30mbps return to most of the State of Alaska. I find that anywhere from 40-65% of our total traffic is "questionable", which is why I was asking about an ISP who liked their users downloading torrents. It is very difficult to gauge a users behavior if they are on an "all out" downloading binge over a weekend. Normally, a user logs in and does what they need to in a relatively short amount of time (hours). In the case of most providers, we oversubscribe our resources and have found this model is beginning to not apply to user behavior changes. Long gone are the days of the user turning off their computers, and our customer base (rural Alaska) have few things to do besides use the internet. This has made a "perfect storm" of sorts, as we are now seeing most of our users utilizing 70%+ of their allocated (purchased) bandwidth 24 hours a day. The vast majority of the night use is gaming, and bit torrent. It makes things much more complicated when trying to give an experience to people..
//warren
-----Original Message----- From: Jack Bates [mailto:jbates@brightok.net] Sent: Tuesday, September 28, 2010 10:26 AM To: Warren Bailey Cc: Richard Barnes; NANOG Subject: Re: Online games stealing your bandwidth
On 9/28/2010 1:00 PM, Warren Bailey wrote:
Jack,
Forgive me if I'm mistaken, but looking at your website - do you only offer dial up services? This could be the background for a statement like "a proper ISP doesn't encourage any type of traffic." We have a couple of OC-192 running to Seattle, so certain "types" of traffic can make a good day turn very badly without some traffic "management".
BrightNet itself has ILEC's as customers. We're a turnkey glue for ILECs nearby. Among other things, I provide engineering support and advise for each ILEC. Each has their own levels of service, management, and technologies deployed including wireless, cellular, DSL, FTTH, and cable. I'm currently running around 1.2gbit between us and 4 NSP transits with 3gbit available. Some of the ILECs have additional load shifting with other transits. I estimate the need to go 10Gig ring or split transit in less than 5 years at current growth rates, and the largest problem we've run into is getting infrastructure to handle gig-e speeds out of rural ILECs for the 100+ mile longhauls. I've had issues with gig-e connectivity just getting out of OKC to enough NSP transits and it will become more difficult/expensive when we do hit 10G.
As it currently stands, we usually have no problems with event spikes, though we sometimes have to tweek the traffic paths depending on how the NSPs do. The largest issues have always been the last mile. As we resolve last mile costs (which dropping 100% fiber in a rural area today doesn't have the safety nets and guarantees that were provided when copper was dropped in), we'll then have to tackle the longhaul connectivity issues, but hopefully the cost to handle that will drop as well.
Jack
What is keeping your company from buying more bandwidth? I find the excuse of over subscription to be a fail. If that's your companies business model then it should not be whining when people are using what you sell them. Provision bandwidth accordingly and stop being cheap and squeezing every last dime from the end user for bandwidth that can be had for less than the price of a burger in some places. Manny -----BEGIN PGP SIGNATURE----- Version: GnuPG/MacGPG2 v2.0.12 (Darwin) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iQEcBAEBAgAGBQJMokBqAAoJEOcnyWxdB1IrGBMH/RCg7zy3L171hwGuilZHRWyA 9B4k+DoTF0Cp8Gt30zamKly90BERKiilryyhxSpAtepUa4wQs4bOGwk5HKx06jkF YJokQpqmNNmY4MU/bwWtUpkjrQjYT6Dt8967iEA3SWBbqdUhRdyejFLaZbDoV43u 61NzEU/JGdxnRvO/MkleP95/+XPCWuQy0EIDAuwlwcWIzr/i9ra9nD5Nf6x9AE/u XTJoTLwY6y2xP93gTBp12MBmzf07AkPxwvpMAbcYIu+94O/twbpWysuceC3EH2bW cMKLPAIROxZaropgSSJYSu8hFNPWlODkOD7MHiY8Ilcv6B4v7XEa6QpCd/lfDxE= =ZPwF -----END PGP SIGNATURE-----
In my experience users aren't willing to pay for dedicated bandwidth. On Tue, Sep 28, 2010 at 1:22 PM, manolo hernandez <mhernand1@comcast.net> wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 9/28/10 3:01 PM, Warren Bailey wrote:
Jack,
Apologies, I did not realize that you guys were doing so much. Please don't take my last email as anything which was intended to question or insult you guys. Up here (Alaska) we have about 100,000 cable subscribers along with mixed Fiber/DSL/POTS access and nearly 50,000 cellular customers with high speed data around our Metro network. I am an RF Engineer, however the network I run is IP based (satellite) and I run in the neighborhood of 250mbps forward and 30mbps return to most of the State of Alaska. I find that anywhere from 40-65% of our total traffic is "questionable", which is why I was asking about an ISP who liked their users downloading torrents. It is very difficult to gauge a users behavior if they are on an "all out" downloading binge over a weekend. Normally, a user logs in and does what they need to in a relatively short amount of time (hours). In the case of most providers, we oversubscribe our resources and have found this model is beginning to not apply to user behavior changes. Long gone are the days of the user turning off their computers, and our customer base (rural Alaska) have few things to do besides use the internet. This has made a "perfect storm" of sorts, as we are now seeing most of our users utilizing 70%+ of their allocated (purchased) bandwidth 24 hours a day. The vast majority of the night use is gaming, and bit torrent. It makes things much more complicated when trying to give an experience to people..
//warren
-----Original Message----- From: Jack Bates [mailto:jbates@brightok.net] Sent: Tuesday, September 28, 2010 10:26 AM To: Warren Bailey Cc: Richard Barnes; NANOG Subject: Re: Online games stealing your bandwidth
On 9/28/2010 1:00 PM, Warren Bailey wrote:
Jack,
Forgive me if I'm mistaken, but looking at your website - do you only offer dial up services? This could be the background for a statement like "a proper ISP doesn't encourage any type of traffic." We have a couple of OC-192 running to Seattle, so certain "types" of traffic can make a good day turn very badly without some traffic "management".
BrightNet itself has ILEC's as customers. We're a turnkey glue for ILECs nearby. Among other things, I provide engineering support and advise for each ILEC. Each has their own levels of service, management, and technologies deployed including wireless, cellular, DSL, FTTH, and cable. I'm currently running around 1.2gbit between us and 4 NSP transits with 3gbit available. Some of the ILECs have additional load shifting with other transits. I estimate the need to go 10Gig ring or split transit in less than 5 years at current growth rates, and the largest problem we've run into is getting infrastructure to handle gig-e speeds out of rural ILECs for the 100+ mile longhauls. I've had issues with gig-e connectivity just getting out of OKC to enough NSP transits and it will become more difficult/expensive when we do hit 10G.
As it currently stands, we usually have no problems with event spikes, though we sometimes have to tweek the traffic paths depending on how the NSPs do. The largest issues have always been the last mile. As we resolve last mile costs (which dropping 100% fiber in a rural area today doesn't have the safety nets and guarantees that were provided when copper was dropped in), we'll then have to tackle the longhaul connectivity issues, but hopefully the cost to handle that will drop as well.
Jack
What is keeping your company from buying more bandwidth? I find the excuse of over subscription to be a fail. If that's your companies business model then it should not be whining when people are using what you sell them. Provision bandwidth accordingly and stop being cheap and squeezing every last dime from the end user for bandwidth that can be had for less than the price of a burger in some places.
Manny -----BEGIN PGP SIGNATURE----- Version: GnuPG/MacGPG2 v2.0.12 (Darwin) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
iQEcBAEBAgAGBQJMokBqAAoJEOcnyWxdB1IrGBMH/RCg7zy3L171hwGuilZHRWyA 9B4k+DoTF0Cp8Gt30zamKly90BERKiilryyhxSpAtepUa4wQs4bOGwk5HKx06jkF YJokQpqmNNmY4MU/bwWtUpkjrQjYT6Dt8967iEA3SWBbqdUhRdyejFLaZbDoV43u 61NzEU/JGdxnRvO/MkleP95/+XPCWuQy0EIDAuwlwcWIzr/i9ra9nD5Nf6x9AE/u XTJoTLwY6y2xP93gTBp12MBmzf07AkPxwvpMAbcYIu+94O/twbpWysuceC3EH2bW cMKLPAIROxZaropgSSJYSu8hFNPWlODkOD7MHiY8Ilcv6B4v7XEa6QpCd/lfDxE= =ZPwF -----END PGP SIGNATURE-----
-----Original Message----- From: Jason Iannone [mailto:jason.iannone@gmail.com] Sent: Tuesday, September 28, 2010 12:50 PM To: manolo hernandez Cc: nanog@nanog.org Subject: Re: Online games stealing your bandwidth
In my experience users aren't willing to pay for dedicated bandwidth.
That's not the point. The point is that if your users are using the net available bandwidth, it's time to add more bandwidth, not to mess with your users' traffic. 'Dedicated' has nothing to do with it. Best Regards, Nathan Eisenberg
On 9/28/2010 2:22 PM, manolo hernandez wrote:
What is keeping your company from buying more bandwidth? I find the excuse of over subscription to be a fail. If that's your companies business model then it should not be whining when people are using what you sell them. Provision bandwidth accordingly and stop being cheap and squeezing every last dime from the end user for bandwidth that can be had for less than the price of a burger in some places.
You replied to him but under my quoted text, so I'm not sure who you were referring to. However, my company has issues in buying long haul. Bandwidth is cheap, yes. Getting a circuit is not. Currently I have 1 option for a 10Gig circuit if I needed it today. That's not very redundant. It took 6 months to get facility upgrades by a large NSP to give me 1gig-e in OKC from DFW (very few NSPs have routers or high speed facilities in Oklahoma and even fewer in OKC. Tulsa has a few extra options). I'm still waiting on what looks like it'll be 1 year+ for a gig-e from another NSP. Going to remote ILEC towns, there's shortages of long haul facilities (in some areas, a single OC-12 sonet run is all that exists and it's dropped off in 3-5 places to various other companies on the way to the ILEC, and the fiber dwindles to 6 meaning primary pair, secondary pair, and backup dark pair is all that exists). The cost to bore new fiber and light it is extremely prohibitive. We actually have no problems with people using what we sell, and we still have nice oversell margins which makes up our profit (0% oversell would be roughly break even). Many of our problems aren't with users using their bandwidth, but with applications screwing with the user's bandwidth (against the user's will). Someone linked bittorrent's work on latency based fallback for congestion control. I think that is an awesome piece of work. However, not all p2p applications do this, and some even install and work in the background without customers knowing. This gives the perception to the customer that things are slow and not working right. We care what our customer's think, so we absolutely hate such products as we can see the bandwidth usage itself, but helping a computer illiterate customer fix the problem without them spending money at a computer tech is difficult at best. Jack
Our excuse? We have purchased every available transponder on every spacecraft suitable for transmission out of Alaska. Granted, there are additional spacecraft out there with Alaska footprints. We however, being a service provider, are interested in space segment which gives us quality over quantity. Sometimes fiber just isn't an option. So that burger analogy doesn't quite apply here .. because the burger is 100mbps, it space segment alone is 150k a month. Not to mention the modems (and remote people who admin them) in the neighborhood of 140k each side of the link. Plus, the diesel used to provide power to the Earth station (9$ a gallon) so it can transmit. Expensive happy meal. -----Original Message----- From: Jack Bates [mailto:jbates@brightok.net] Sent: Tuesday, September 28, 2010 12:20 PM To: manolo hernandez Cc: nanog@nanog.org Subject: Re: Online games stealing your bandwidth On 9/28/2010 2:22 PM, manolo hernandez wrote:
What is keeping your company from buying more bandwidth? I find the excuse of over subscription to be a fail. If that's your companies business model then it should not be whining when people are using what you sell them. Provision bandwidth accordingly and stop being cheap and squeezing every last dime from the end user for bandwidth that can be had for less than the price of a burger in some places.
You replied to him but under my quoted text, so I'm not sure who you were referring to. However, my company has issues in buying long haul. Bandwidth is cheap, yes. Getting a circuit is not. Currently I have 1 option for a 10Gig circuit if I needed it today. That's not very redundant. It took 6 months to get facility upgrades by a large NSP to give me 1gig-e in OKC from DFW (very few NSPs have routers or high speed facilities in Oklahoma and even fewer in OKC. Tulsa has a few extra options). I'm still waiting on what looks like it'll be 1 year+ for a gig-e from another NSP. Going to remote ILEC towns, there's shortages of long haul facilities (in some areas, a single OC-12 sonet run is all that exists and it's dropped off in 3-5 places to various other companies on the way to the ILEC, and the fiber dwindles to 6 meaning primary pair, secondary pair, and backup dark pair is all that exists). The cost to bore new fiber and light it is extremely prohibitive. We actually have no problems with people using what we sell, and we still have nice oversell margins which makes up our profit (0% oversell would be roughly break even). Many of our problems aren't with users using their bandwidth, but with applications screwing with the user's bandwidth (against the user's will). Someone linked bittorrent's work on latency based fallback for congestion control. I think that is an awesome piece of work. However, not all p2p applications do this, and some even install and work in the background without customers knowing. This gives the perception to the customer that things are slow and not working right. We care what our customer's think, so we absolutely hate such products as we can see the bandwidth usage itself, but helping a computer illiterate customer fix the problem without them spending money at a computer tech is difficult at best. Jack
participants (6)
-
Jack Bates -
Jason Iannone -
manolo hernandez -
Nathan Eisenberg -
Richard Barnes -
Warren Bailey