It's also an open secret that a number of Google Advertisers have had their accounts suspended and payments withheld because of "Fraudulent Clicks" on a website. Google refuse to disclose any details of what they think is causing the issue when this happens - I've been warned by Google about "violations of the Acceptable Use Policy" with absolutely no other detail as to what I'm supposed to have done. Any queries are met with canned replies. (They would not actually be able to get away with this in the UK or many other European countires due to the Data Protection Act and similar - they can be forced to give up any information they hold)
They are very much throwing the baby out with the bathwater -- it's perfectly possible to kill a rivals cash flow if they're using Google simply by running a bot to click on all the ads on their site. (I think this is what happened in my case) Of course, as Google present no evidence you can't then sue your rival.
I would immediately switch to some other advertising network if there was one available for smaller (~8-9 million hits a month) web sites in the UK. Sadly, there isn't - yet.
"If it takes three minutes for a response, it is not really HTTP any more," Box said. The problem, said Box, is that the intermediaries--that is, the companies that own the routers and cables between the client and server--will not allow single transactions that take this long.
The "intermediaries" only talk IP, they don't parse the TCP or HTTP. IP is stateless - you've no idea when you route/switch a packet if it's from a pre-exsiting connection or a new one as you only look at the source and destination IPs. It's irrelevant if the "intermediaries" have low TCP/HTTP timeouts set because you're not talking TCP/HTTP to those devices. As long as the 'net is still passing traffic the TCP connection will never time out between the browser and the server - as long as the user doesn't hit "stop" - because TCP sends keepalives.
Either the.net engineer is not very clueful (Unlikely) or ZDNet are misreporting.
Token Ring has a lot of features that Ethernet doesn't, such as Layer 2 load balancing. Ethernet is just as susceptable to problems as Token Ring, but most large corporates using either have moved to switching technology by now which tends to isolate problems to single machines. It's not as easy for many large corporates to switch technologies as some people suggest, they probably have significant investments in staff training, spares stock and network infrastructure as well as just the workstation NICs. If they are an IBM house, they may also rely heavily on Source Route Bridging to access servers and switching to Ethernet would require all new software on servers and clients and possibly a major reengineer of their main network.
On the topic of the "Ring errors", more information is needed. Exactly what errors were being seen, and is the end station connected to a dumb MAU or a modern TR switch. I suspect for some reason the Linux workstation is connected to an old MAU (Like an ethernet hub or even old 10Base-2 cabling) rather than a switch and for some reason the card is repeatedly resetting or dropping the token, which is probably a driver fault given the fact it works in Windows. There's not necessarily a lot that could have been done to pick this up, because it doesn't immediately bring the ring down and very few corporates really stress-test every possible piece of hardware, software and driver before deploying it unless it's part of a major rollout of a new technology. Had the same problem occured in the Windows driver as part of an update, I doubt it would have been picked up before deployment unless part of a major rollout for example.
It's not a case of "removing" the filtering, you'd have to build something to either specifically pass through a certain signal or build a sort of inverse notch filter which filters everything but a certain frequency. This isn't easy - filtering is basically done by bunging a very high AC capacatance on the line to smooth out any spikes or sags.
I don't think their technology is based on this however, I think they've found something that's immune to "normal" filtering.
Southern District Magistrate Judge Henry Pitman... ruled that AOL was entitled to recover for unjust enrichment, since Christian Brothers unlawfully used the AOL mark and misappropriated services that otherwise could have been sold to advertisers.
The way I read that, AOL are complaining that they wanted to spam their own customers and got damages on the basis of lost spam-generated revenue. I'm not sure what they mean about "misappropriating the AOL mark". Perhaps the Christian Brothers are one of the group of spammers that use somejunkaddress@aol.com as the From: field.
A quick troll through the rather vague Media Fusion web site suggests that the signal is one-way, at least insofar as they've managed to solve the "bypassing the transformer" problem. I can't see any indication that the user can send (large amounts of) data. Maybe just a cheap version of Cable TV rather than a cheap Internet?
This is not about sending an electric signal over power lines. This is about manipulating the magnetic field which forms naturaly around a wire when you send an A/C current through it. Sounds like just marketing misunderstanding the technology. You can't manipulate the wave surrounding the current without affecting the current. They may however use the wave to read from or modify the current which might be more sensitive than manipulating it directly.
IIRC, they tried this to provide local loop services but found that they were turning the streetlamps into broadcasting radio antennas. I believe the authorities stamped on it pretty quick when they found out. It's not an insurmountable problem but I think they found that it wasn't worth the cost.
Haven't we seen this before? Usually it turns out to be nothing new - either the old trick of using the earth wire for communciations (Which "baby monitors" already do and have done for years) or just electric companies selling pylon space for fibre. This is already done in the UK by Energis.
You can't provide long-distance communications over electric cabling because Live and Neutral are filtered at the substations and would remove any meaningful signal and the earth wire is local to each house. You could *possibly* set up local loops this way but you don't gain anything - it's the long-haul services and hardware that cost.
Of course, copying money wouldn't be a problem if the US actually used some of the methods available this side of the pond. For example, English currency has a distinct feel to the paper it and as well as the usual metal strip and watermark has "raised" ink that rubs off slighly. (Try rubbing the "Bank of England" on the queens head side of a newish note on some light coloured trousers) Higher value notes even have holigrams. The Germans go even further - they have UV ink, perforations in the notes and other features. Even official documents like the V5 (Vechicle Registration Document) have anti-copy features. I recently photocopied mine in case I lost the original (Which I usually carry on me) and was quite surprised to find that the patterned background of the document had "COPY" quite clearly all over it. Close examination revealed thay you can just make out the "COPY" watermark on the original.
Yet another case of attacking the problem from the wrong end - the US are trying to limit the technology instead of staying one step ahead of the game.
Slashdot Mirror
Rather a lot of us in fact.
It's also an open secret that a number of Google Advertisers have had their accounts suspended and payments withheld because of "Fraudulent Clicks" on a website. Google refuse to disclose any details of what they think is causing the issue when this happens - I've been warned by Google about "violations of the Acceptable Use Policy" with absolutely no other detail as to what I'm supposed to have done. Any queries are met with canned replies. (They would not actually be able to get away with this in the UK or many other European countires due to the Data Protection Act and similar - they can be forced to give up any information they hold)
They are very much throwing the baby out with the bathwater -- it's perfectly possible to kill a rivals cash flow if they're using Google simply by running a bot to click on all the ads on their site. (I think this is what happened in my case) Of course, as Google present no evidence you can't then sue your rival.
I would immediately switch to some other advertising network if there was one available for smaller (~8-9 million hits a month) web sites in the UK. Sadly, there isn't - yet.
The "intermediaries" only talk IP, they don't parse the TCP or HTTP. IP is stateless - you've no idea when you route/switch a packet if it's from a pre-exsiting connection or a new one as you only look at the source and destination IPs. It's irrelevant if the "intermediaries" have low TCP/HTTP timeouts set because you're not talking TCP/HTTP to those devices. As long as the 'net is still passing traffic the TCP connection will never time out between the browser and the server - as long as the user doesn't hit "stop" - because TCP sends keepalives.
Either the
Token Ring has a lot of features that Ethernet doesn't, such as Layer 2 load balancing. Ethernet is just as susceptable to problems as Token Ring, but most large corporates using either have moved to switching technology by now which tends to isolate problems to single machines. It's not as easy for many large corporates to switch technologies as some people suggest, they probably have significant investments in staff training, spares stock and network infrastructure as well as just the workstation NICs. If they are an IBM house, they may also rely heavily on Source Route Bridging to access servers and switching to Ethernet would require all new software on servers and clients and possibly a major reengineer of their main network.
On the topic of the "Ring errors", more information is needed. Exactly what errors were being seen, and is the end station connected to a dumb MAU or a modern TR switch. I suspect for some reason the Linux workstation is connected to an old MAU (Like an ethernet hub or even old 10Base-2 cabling) rather than a switch and for some reason the card is repeatedly resetting or dropping the token, which is probably a driver fault given the fact it works in Windows. There's not necessarily a lot that could have been done to pick this up, because it doesn't immediately bring the ring down and very few corporates really stress-test every possible piece of hardware, software and driver before deploying it unless it's part of a major rollout of a new technology. Had the same problem occured in the Windows driver as part of an update, I doubt it would have been picked up before deployment unless part of a major rollout for example.
It's not a case of "removing" the filtering, you'd have to build something to either specifically pass through a certain signal or build a sort of inverse notch filter which filters everything but a certain frequency. This isn't easy - filtering is basically done by bunging a very high AC capacatance on the line to smooth out any spikes or sags.
I don't think their technology is based on this however, I think they've found something that's immune to "normal" filtering.
The way I read that, AOL are complaining that they wanted to spam their own customers and got damages on the basis of lost spam-generated revenue. I'm not sure what they mean about "misappropriating the AOL mark". Perhaps the Christian Brothers are one of the group of spammers that use somejunkaddress@aol.com as the From: field.
A quick troll through the rather vague Media Fusion web site suggests that the signal is one-way, at least insofar as they've managed to solve the "bypassing the transformer" problem. I can't see any indication that the user can send (large amounts of) data. Maybe just a cheap version of Cable TV rather than a cheap Internet?
This is not about sending an electric signal over power lines. This is about manipulating the magnetic field which forms naturaly around a wire when you send an A/C current through it.
Sounds like just marketing misunderstanding the technology. You can't manipulate the wave surrounding the current without affecting the current. They may however use the wave to read from or modify the current which might be more sensitive than manipulating it directly.
IIRC, they tried this to provide local loop services but found that they were turning the streetlamps into broadcasting radio antennas. I believe the authorities stamped on it pretty quick when they found out. It's not an insurmountable problem but I think they found that it wasn't worth the cost.
Haven't we seen this before? Usually it turns out to be nothing new - either the old trick of using the earth wire for communciations (Which "baby monitors" already do and have done for years) or just electric companies selling pylon space for fibre. This is already done in the UK by Energis.
You can't provide long-distance communications over electric cabling because Live and Neutral are filtered at the substations and would remove any meaningful signal and the earth wire is local to each house. You could *possibly* set up local loops this way but you don't gain anything - it's the long-haul services and hardware that cost.
Of course, copying money wouldn't be a problem if the US actually used some of the methods available this side of the pond. For example, English currency has a distinct feel to the paper it and as well as the usual metal strip and watermark has "raised" ink that rubs off slighly. (Try rubbing the "Bank of England" on the queens head side of a newish note on some light coloured trousers) Higher value notes even have holigrams. The Germans go even further - they have UV ink, perforations in the notes and other features. Even official documents like the V5 (Vechicle Registration Document) have anti-copy features. I recently photocopied mine in case I lost the original (Which I usually carry on me) and was quite surprised to find that the patterned background of the document had "COPY" quite clearly all over it. Close examination revealed thay you can just make out the "COPY" watermark on the original.
Yet another case of attacking the problem from the wrong end - the US are trying to limit the technology instead of staying one step ahead of the game.