Well, you have to understand that traceroutes are replied by routers, and routers might choose to prioritise ICMP (the protocol used by traceroute replies and ping) lower than normal traffic. So you should only use traceroutes to get an idea of the path the data takes, not the health of that path.
I'm a network engineer at an ISP, so I would say I have a bit of experience with this from both ends of the table. First of all, there's a difference between your broadband connection speed and your perceived rate. Your broadband connection might be capped to what you pay for, and, assuming your last-mile medium can handle that speed, that only means that you will never actually go beyond your connection speed.
Now as we know, the internet is a complicated network of interconnected systems. You are connected via your ISP's backbone to the other systems (ISPs, enterprises, content providers, etc.) via a number of internet peering points. These peering points have their own connection speed (typically 1 Gbit/s or 10 Gbit/s, although higher exist), and may or may not be utilised to their maximum extent at any point of time. This means that you may have your full data rate available to some destinations, while others may take a congested route.
You mention testing, and your frustration is very reasonable. There are testing sites out there, but you never have any idea about how many else might be testing at the same time, or how much load there is on the server at the moment of the test. If you are unlucky, you might also be limited by your hardware, your operating system (TCP Window Size, receive buffers and similar might not be tuned properly), or your router.
I would say your best choice would be to download as much as possible from as many sources as possible (bittorrent is excellent for this, but may be throttled by evil ISPs), and do this over a couple of days to get an average indication of how much your connection is capable of delivering.
If you have a server on some remote location via the internet, you can use programs like iperf to make a bandwidth test, but such a test is not exactly precise when you have no idea how the intermediate networks are.
In short, it's just too early to tell. Just because the RIRs ran out of addresses, it doesn't mean that the LIRs have yet (the ISPs).
Based on my experience as a network engineer at an ISP, the following is happening already:
Small ISPs and ISPs that have not been in the business for a long time* have either run out or are on the verge of doing so. They are doing the following:
* Purchasing legacy IPv4 addresses from enterprises with/16 networks from the old days where available.
* Deploying CGN-like solutions for their end-customers if their end-customers are residential users.
Larger ISPs and older ISPs with allocations from ye old pre-RIR days continue to hold addresses and are often able to free large quantities of addresses from old deployments. Mind you, a lot of public IPv4 space have been "wasted" on infrastructure addressing, and management of devices that were not even connected to the internet. Devices such as modems, DSLAMs, CPEs and similar.
One could easily speculate that the business of ISPs will be severely affected in the future, as customers will go to the old providers that have plenty of v4-space available at the cost of newer players who followed the RIR regulations of only applying for the address space they needed based on relative short-term predictions.
If you are a registered LIR you will see a flood of SPAM from so-called IP brokers who are trying to purchase unused IPv4 space in hope of selling this to LIRs in need. That market will probably become quite desperate in the coming years.
Oh, and by the way, I see no evidence that IPv6 deployment is taking any noticeable speed.
*) Long as in they were in the game when classfull allocations were made.
Well, if by "run out" you mean have no addresses left at all, you are correct. Subdividing the number of LIRs by the number of possible/22 in the remaining/8, they will not run out, but once you get your final/22, there are no more IPv4 addresses to get. The remaining addresses are yet to have their purpose declared, and I cannot imagine that they will be allocated in any way resembling current address policies. Within RIPE policy, the remaining addresses in the final/8 are reserved for unforeseen events, whatever that may be...
APNIC is the fastest growing RIR out there, with developing markets like China and India burning IP addresses like crazy. These markets will grow beyond what we have seen already, simply because more people are getting internet access daily.
Sure, you can do carrier-grade NAT, Large Scale NAT and whatever the appropriate name of ISP-level NAT is these days, but for a market of the size of China, a/22 only gets you so far.
According to data from the CNNIC - China Network Information Center - the number of broadband users in China is currently around 457 million. In very round numbers, that gives us only a third of the entire population, and with a growth of 48% per year, it's very obvious why a/22 doesn't last very long at all.
Before you even mention it, the/22 is per LIR, but there are not that many LIRs in China due central government control.
Online business in China is striving too, and you just can't build up a website without routable addresses.
Interestingly enough I saw exactly the same. Two of our production servers running fedora 8 crashed exactly at 01:01:02 GMT+1.
I am beginning to suspect that this must be a Fedora 8/NTP-related bug...
That sounds awfully a lot like the old ISO addressing. They were terribly inconvenient, but are still used for a lot of telecom equipment (SDH/SONET is a good example).
I once worked in a call center, where we used a pc soft phone that would pop up and take focus when ever a customer called. We usually answered by pressing spacebar or enter, but whenever windows update popped up, you quite often ended up being unable to help the customer as you accidentally hit "Reboot now" and your PC started rebooting just when the customer came through to you.
This is understandable, considering the level of obnoxiousness. Firefox gives you a discrete notice that it has downloaded an update, and you can choose not to install it right away, but instead having it installed next time you start firefox. Windows Updates are so damn obnoxious that I always consider turning it off and doing my updates manually. I know how to update my computer manually, but I suspect the bulk of users out there, just get frustrated about the constant bells and whistles of Windows Update, that they turn it off and leave it like that.
And according to this link (no myminicity, I swear!), Helium is in danger of being in short supply due to among other things that it's not captured and recycled after use and while being available in big supply in the universe, the Earth supply is actually a bit limited.
According to the article it is an issue the next generations of scientist are going to have to struggle with. So maybe a Helium-based airship is not that good an idea, although I don't have to background to propose a different scheme.
Quite true. And I'm very happy that I'm not currently sitting on an X.25 network transfering files over FTAM, meanwhile I try to converse over x.400...
Back to the topic: I had quite good experience with the IPv6 autoconfiguration protocol, but I must recognice the point of the article. Autoconfiguration works quite well when you need to tell your clients what/64 they are on, but it stops there. There will always be a demand to specify additional information like specify name servers, tftp, ntp etc.
Well, neither can an electric RJ45. When running at those speed, optical connection (fiber) is necessary. GbE is still available as electric (and point-to-point wireless), but when running at higher speeds like 10GbE, you can't continue using your RJ45 cable, not that it's a problem anyway, it's mostly carriers/ISPs who use 10GbE and for those, optical cables aren't an issue at all.
That is probably the closest to the truth, however I keep wondering if, by releasing the full details of a bug or security hole to the public, you force the developers to make patches that fixes that specific exploit but leaves open the hole, thus protecting the software from script kiddies browsing through security sites in search of exploits, but not from being cracked again in the future. What if the company did have some truth in saying that the bug could only be thoroughly fixed by rewriting the software from scratch?
I'm not saying that this is the truth in every case at all, but I think it is worth considering in this debate.
Working for an isp that primarily focuses on wireless technology, I am looking forward to this. Although I've read some valid points for those of you who focus on the data center part of this, there is another line of bussiness where this actually is a good idea. First of all, fibre being expensive to dig down, takes a long time to install and has all the obvious obstacles and problems (it has obvious advantages too, no doubt about it). Here my company has had success in using STM-1 (or OC-3 for you Americans) radiolinks converted into Ethernet, or even faster radiolinks now available for back haul to places where fibre wasn't an option (or at least a too expensive option). A radiolink can be installed in a day, where as fibre might take a month or two. Having a wireless connection that can perform more than a gigabit per second can only be an advantage, and can push high-speed internet connections to areas where this was previously impossible.
Another aspect of this is the dominance of copper-based (DSL) connections that in many countries are monopolized through one company owning the entire copper net, and furthermore are limited in their transmission capabilities. Although technologies such as VDSL and VDSL2 promisses up to 100Mbit synchronous, they have a lot of issues regarding the requirements to the quality of the copper, the maximum distance to the DSLAM etc. (Wikipedia on VDSL2)
High-speed wireless connection could change this dominance and bring true synchronous broadband to the people. While I have no illusions that your average John and Jane Doe will have a 60Ghz wireless connection on their roof top, it will ease the roll out of other wireless connections, for instance the much-hyped WiMAX technology, and enable cariers to provide a "holistic" wireless service with all the beforementioned advantages.
You still have to consider the backhaul isn't fibre, or very rarely is fibre. Most likely it's wireless radio link, such as Ericsson Mini-Link, which has a total capacity of 34mbit/s.
I don't work in mobile telephone communications, but my bussiness has quite a lot of simmilarities. if this indeed is 3G data, I can very well understand why they want to restrict the usage, as the data is caried over the same backhaul as the GSM cell phone conversations (IP data encapsulated in a DS3 or E3 radiolink backhaul most likely). As GSM conversations are transported via SONET/SDH, they retain a constant bitrate, and therefore leaves little space left for data transport, which has to "fight" for the allocated bandwidth (just like an ATM UBR type), thus if somebody hooked up their 3G cellphone (or PCMCIA card for that matter) to their computer, in a very short distance to their GSM base, they could very well overload the entire base station. So in that case, having limits to a mobile internet service makes indeed sense, calling it unlimited however does not, and I agree with everyone that this kind of advertisement should be banned.
That said, I wouldn't mind this kind of subscription. My current cell phone plan includes a whooping 10MB of free trafic every month and an extra 10 DKR (approx. $1.79).
Well, yes, you are right. The transmit power of a wimax connection, or any other network connection in that spectrum, is not nearly strong enough to interfere with any airline systems. It would probably be the other way around I presume:)
Well, 802.11b/g doesn't mess up the navigational systems, but 802.11a does, if you have any of that legacy equipment. It uses 5.5Ghz which is also the spectre in which radar is broadcasted.
This is also the frequency many unlicensed WiMAX connections used, although primarily in countries where the original 3.5ghz spectrum is not available for licensing.
There is a certain list of "bad words" that you are not allowed to use in email, and that you WILL get into trouble with if you use them. Also sending an email to a Chinese from anywhere containing one or more of these words have a high chance of getting the receiver into trouble.
it would certainly beat out my first 8 years of education, where I've found almost all of the science I've learned is actually wrong (by talking to scientists, and reading books, and wikipedia).
You wouldn't happen to be from Kansas, would you?;)
Slashdot Mirror
Well, you have to understand that traceroutes are replied by routers, and routers might choose to prioritise ICMP (the protocol used by traceroute replies and ping) lower than normal traffic. So you should only use traceroutes to get an idea of the path the data takes, not the health of that path.
I'm a network engineer at an ISP, so I would say I have a bit of experience with this from both ends of the table. First of all, there's a difference between your broadband connection speed and your perceived rate. Your broadband connection might be capped to what you pay for, and, assuming your last-mile medium can handle that speed, that only means that you will never actually go beyond your connection speed.
Now as we know, the internet is a complicated network of interconnected systems. You are connected via your ISP's backbone to the other systems (ISPs, enterprises, content providers, etc.) via a number of internet peering points. These peering points have their own connection speed (typically 1 Gbit/s or 10 Gbit/s, although higher exist), and may or may not be utilised to their maximum extent at any point of time. This means that you may have your full data rate available to some destinations, while others may take a congested route.
You mention testing, and your frustration is very reasonable. There are testing sites out there, but you never have any idea about how many else might be testing at the same time, or how much load there is on the server at the moment of the test. If you are unlucky, you might also be limited by your hardware, your operating system (TCP Window Size, receive buffers and similar might not be tuned properly), or your router.
I would say your best choice would be to download as much as possible from as many sources as possible (bittorrent is excellent for this, but may be throttled by evil ISPs), and do this over a couple of days to get an average indication of how much your connection is capable of delivering.
If you have a server on some remote location via the internet, you can use programs like iperf to make a bandwidth test, but such a test is not exactly precise when you have no idea how the intermediate networks are.
In short, it's just too early to tell. Just because the RIRs ran out of addresses, it doesn't mean that the LIRs have yet (the ISPs).
Based on my experience as a network engineer at an ISP, the following is happening already:
Small ISPs and ISPs that have not been in the business for a long time* have either run out or are on the verge of doing so. They are doing the following:
* Purchasing legacy IPv4 addresses from enterprises with /16 networks from the old days where available.
* Deploying CGN-like solutions for their end-customers if their end-customers are residential users.
Larger ISPs and older ISPs with allocations from ye old pre-RIR days continue to hold addresses and are often able to free large quantities of addresses from old deployments. Mind you, a lot of public IPv4 space have been "wasted" on infrastructure addressing, and management of devices that were not even connected to the internet. Devices such as modems, DSLAMs, CPEs and similar.
One could easily speculate that the business of ISPs will be severely affected in the future, as customers will go to the old providers that have plenty of v4-space available at the cost of newer players who followed the RIR regulations of only applying for the address space they needed based on relative short-term predictions.
If you are a registered LIR you will see a flood of SPAM from so-called IP brokers who are trying to purchase unused IPv4 space in hope of selling this to LIRs in need. That market will probably become quite desperate in the coming years.
Oh, and by the way, I see no evidence that IPv6 deployment is taking any noticeable speed.
*) Long as in they were in the game when classfull allocations were made.
Well, if by "run out" you mean have no addresses left at all, you are correct. Subdividing the number of LIRs by the number of possible /22 in the remaining /8, they will not run out, but once you get your final /22, there are no more IPv4 addresses to get. The remaining addresses are yet to have their purpose declared, and I cannot imagine that they will be allocated in any way resembling current address policies. Within RIPE policy, the remaining addresses in the final /8 are reserved for unforeseen events, whatever that may be...
I call bullshit on your bullshit.
APNIC is the fastest growing RIR out there, with developing markets like China and India burning IP addresses like crazy. These markets will grow beyond what we have seen already, simply because more people are getting internet access daily.
Sure, you can do carrier-grade NAT, Large Scale NAT and whatever the appropriate name of ISP-level NAT is these days, but for a market of the size of China, a /22 only gets you so far.
According to data from the CNNIC - China Network Information Center - the number of broadband users in China is currently around 457 million. In very round numbers, that gives us only a third of the entire population, and with a growth of 48% per year, it's very obvious why a /22 doesn't last very long at all.
Before you even mention it, the /22 is per LIR, but there are not that many LIRs in China due central government control.
Online business in China is striving too, and you just can't build up a website without routable addresses.
Everyone knows that a bad analogy is like a leaking screwdriver...
Interestingly enough I saw exactly the same. Two of our production servers running fedora 8 crashed exactly at 01:01:02 GMT+1. I am beginning to suspect that this must be a Fedora 8/NTP-related bug...
That sounds awfully a lot like the old ISO addressing. They were terribly inconvenient, but are still used for a lot of telecom equipment (SDH/SONET is a good example).
I once worked in a call center, where we used a pc soft phone that would pop up and take focus when ever a customer called. We usually answered by pressing spacebar or enter, but whenever windows update popped up, you quite often ended up being unable to help the customer as you accidentally hit "Reboot now" and your PC started rebooting just when the customer came through to you.
This is understandable, considering the level of obnoxiousness. Firefox gives you a discrete notice that it has downloaded an update, and you can choose not to install it right away, but instead having it installed next time you start firefox. Windows Updates are so damn obnoxious that I always consider turning it off and doing my updates manually. I know how to update my computer manually, but I suspect the bulk of users out there, just get frustrated about the constant bells and whistles of Windows Update, that they turn it off and leave it like that.
This would have been the first post, if I remembered to set my clock for DST.
And according to this link (no myminicity, I swear!), Helium is in danger of being in short supply due to among other things that it's not captured and recycled after use and while being available in big supply in the universe, the Earth supply is actually a bit limited.
According to the article it is an issue the next generations of scientist are going to have to struggle with. So maybe a Helium-based airship is not that good an idea, although I don't have to background to propose a different scheme.
I, for one, welcome our new glow in the dark cat overlords!
Sorry, couldn't help it...
Quite true. And I'm very happy that I'm not currently sitting on an X.25 network transfering files over FTAM, meanwhile I try to converse over x.400... Back to the topic: I had quite good experience with the IPv6 autoconfiguration protocol, but I must recognice the point of the article. Autoconfiguration works quite well when you need to tell your clients what /64 they are on, but it stops there. There will always be a demand to specify additional information like specify name servers, tftp, ntp etc.
Well, neither can an electric RJ45. When running at those speed, optical connection (fiber) is necessary. GbE is still available as electric (and point-to-point wireless), but when running at higher speeds like 10GbE, you can't continue using your RJ45 cable, not that it's a problem anyway, it's mostly carriers/ISPs who use 10GbE and for those, optical cables aren't an issue at all.
That is probably the closest to the truth, however I keep wondering if, by releasing the full details of a bug or security hole to the public, you force the developers to make patches that fixes that specific exploit but leaves open the hole, thus protecting the software from script kiddies browsing through security sites in search of exploits, but not from being cracked again in the future. What if the company did have some truth in saying that the bug could only be thoroughly fixed by rewriting the software from scratch? I'm not saying that this is the truth in every case at all, but I think it is worth considering in this debate.
In soviet russia, the internet destroys Elton John...
Working for an isp that primarily focuses on wireless technology, I am looking forward to this. Although I've read some valid points for those of you who focus on the data center part of this, there is another line of bussiness where this actually is a good idea. First of all, fibre being expensive to dig down, takes a long time to install and has all the obvious obstacles and problems (it has obvious advantages too, no doubt about it). Here my company has had success in using STM-1 (or OC-3 for you Americans) radiolinks converted into Ethernet, or even faster radiolinks now available for back haul to places where fibre wasn't an option (or at least a too expensive option). A radiolink can be installed in a day, where as fibre might take a month or two. Having a wireless connection that can perform more than a gigabit per second can only be an advantage, and can push high-speed internet connections to areas where this was previously impossible.
Another aspect of this is the dominance of copper-based (DSL) connections that in many countries are monopolized through one company owning the entire copper net, and furthermore are limited in their transmission capabilities. Although technologies such as VDSL and VDSL2 promisses up to 100Mbit synchronous, they have a lot of issues regarding the requirements to the quality of the copper, the maximum distance to the DSLAM etc. (Wikipedia on VDSL2)
High-speed wireless connection could change this dominance and bring true synchronous broadband to the people. While I have no illusions that your average John and Jane Doe will have a 60Ghz wireless connection on their roof top, it will ease the roll out of other wireless connections, for instance the much-hyped WiMAX technology, and enable cariers to provide a "holistic" wireless service with all the beforementioned advantages.
You still have to consider the backhaul isn't fibre, or very rarely is fibre. Most likely it's wireless radio link, such as Ericsson Mini-Link, which has a total capacity of 34mbit/s.
I don't work in mobile telephone communications, but my bussiness has quite a lot of simmilarities. if this indeed is 3G data, I can very well understand why they want to restrict the usage, as the data is caried over the same backhaul as the GSM cell phone conversations (IP data encapsulated in a DS3 or E3 radiolink backhaul most likely). As GSM conversations are transported via SONET/SDH, they retain a constant bitrate, and therefore leaves little space left for data transport, which has to "fight" for the allocated bandwidth (just like an ATM UBR type), thus if somebody hooked up their 3G cellphone (or PCMCIA card for that matter) to their computer, in a very short distance to their GSM base, they could very well overload the entire base station. So in that case, having limits to a mobile internet service makes indeed sense, calling it unlimited however does not, and I agree with everyone that this kind of advertisement should be banned.
That said, I wouldn't mind this kind of subscription. My current cell phone plan includes a whooping 10MB of free trafic every month and an extra 10 DKR (approx. $1.79).
Well, yes, you are right. The transmit power of a wimax connection, or any other network connection in that spectrum, is not nearly strong enough to interfere with any airline systems. It would probably be the other way around I presume :)
Well, 802.11b/g doesn't mess up the navigational systems, but 802.11a does, if you have any of that legacy equipment. It uses 5.5Ghz which is also the spectre in which radar is broadcasted. This is also the frequency many unlicensed WiMAX connections used, although primarily in countries where the original 3.5ghz spectrum is not available for licensing.
Will this mean that I will never be able to search for "Girls with breasts bigger than 36D"?
Actually, this is not far from how it works.
There is a certain list of "bad words" that you are not allowed to use in email, and that you WILL get into trouble with if you use them. Also sending an email to a Chinese from anywhere containing one or more of these words have a high chance of getting the receiver into trouble.
it would certainly beat out my first 8 years of education, where I've found almost all of the science I've learned is actually wrong (by talking to scientists, and reading books, and wikipedia). You wouldn't happen to be from Kansas, would you? ;)