You are just nitpicking. In your original post you are showing ignorance about effective radiated power by talking about "just 0.25 watts", making me try to explain that issue to the layman, and then you suddenly creep out under the cover and claim you are an extra class amateur radio operator. In that case you should have known that communication over thousands of kilometers have been made on 2.4 GHz and your claim about the power being too low is totally irrelevant.
Before turning down my calculation you should make your own math with two 2.1 dB antennas when compared with two 20 dB antennas and 0.25 W of transmitting power and you will see that with the 2.1 dB antennas you will need close to 1000 Watts of power to have the same effective signal. So what is your point?
Line of sight has never been a topic of the discussion, it was only about range and required output power. Furthermore, I don't think you can run an 11 Mbps link or even the link speed they achieved in this case on HF bands with 5 watts.
But that requires another idiot: the one that allows employees to take unprotected systems home, connect them to other networks, and then back on the company LAN.
How did it enter your very large company? I cannot believe that companies have port 445 open from the outside, so it must have entered a different way.
Badly firewalled laptops that pick it up while connected to the wireless internet at home and then release it back on the LAN?
You forget that they used 2ft dish antennas. An antenna like this would have about 20dB of gain. That amplifies the 0.25 watts to 25 watts! And, it also amplifies the received signal by the same amount. When compared with an HF bands signal transmitted by a typical dipole antenna, you would have the equivalent of about 1000 watts of power, assuming a setup like shown in the picture is used at both ends. Of course you need line-of-sight, but the station apparently sits on a hilltop.
When comparing HF to SHF, people often don't take this factor in the calculation. In fact, when calculating the total of achievable antenna gain within a reasonable size, path loss, and achievable transmission power using off-the-shelf technology, the 2.4 GHz band is about the optimum band for getting long range line-of-sight communication. Amateur satellites have been built that transmit only a couple of watts and are (were) easily receivable over a distance of 40000 km.
Connect a decent antenna. Bundling of the radio energy is the key to covering very long distances with very small transmitted power. And it also works on the receiving end, so you can start doing this at one end and see if it helps. A moderately directive antenna will amplify the signal by 50 times at this frequency, and is commercially available for $100 or so.
I don't get it... what is impressive about transmitting in the 2.4 GHz range? This may have been a feat 4 decades ago, but today the parts to construct a transmitter operating in that range are commodity. Shown by the fact that anyone can buy a digital wireless transmission system for under 200 dollars.
I don't think Cisco belongs in that list. Cisco stuff is mainly overpriced hardware running an outdated monolytic piece of software that has too many bolted on additions and features. It is more or less like Windows, in that regard. And they market it marketed so well that people still think it is the best and that it is worth the money. But in reality the company will not even support the customer who bought their expensive hardware; they will have to pay separately for support. And small but essential feature updates cannot be made easily because everything has to go in the generally released monolythic IOS.
Cisco IOS is like running a system like Knoppix on all your production machines. You can make changes to some pre-determined configuration item but not to any part of the functionality.
Not only will it be difficult to form a public opinion on issues such as this, but also it is a widespread misbelief that public opinion has any direct effect on laws and international treaties. Especially when patents, intellectual property and similar issues are involved, the indirect democratic system that most countries have means that the public has no influence. You get a chance to vote a politician, yes. But people in politics to not tend to be that revolting once they are in placs like parliament. And international business has a very effective lobbying system in place, which convinces those politicians that it is best for their countries economy to favour multinational big business. Patents and intellectual property laws are part of that.
"voicing out loudly" is not going to make a better impression than thick reports about profits and jobs, and the occasional bribe.
One shortcoming of current word processing applications programs in particular is exposed when a user attempts to locate numerical information in a text document utilizing a standard "find" function.
This I noticed 20 years ago, but apparently Microsoft has now noticed it as well.
Free tip: Many Unix text editors had powerful features for searching complicated patterns even before "current word processing applications programs" were even developed. So, instead of filing yet another stupid patent, you could set a team to work and implement a decent regexp search and replace.
This usually does not help much, because on top of the TCP connection there usually exists some application protocol that sends a request, waits for an answer, and then sends the next request.
For example, if you would use such a link to provide a VPN between your home network and a network at your employer, and you would use this to logon to that network with a Windows system, you would need to wait at least 10 minutes for the logon to complete even with locally stored profile etc. This is because Windows sends thousands of requests over the TCP connection(s) between client and server, and awaits a response on each of them before proceeding.
And then we are not even talking about thin client computing...
The current offering from Hughes, DIRECWAY, has over 250,000 broadband customers.
250,000?? In this week's news, I read that the major ADSL provider in our small country (which has 15 million inhabitants), has 1.6 million subscribers. And there are other ADSL providers (together something like 400,000 subscribers) and Cable (with over 1 million subscribers).
Internet via satellite cannot make even a small dent into that. And of course there are many reasons for that.
Antenna gain and beamwidth can be directly derived from the antenna size, when you know the frequency it operates at. For 66ft at GigaHz frequencies you will have a lot of gain and a small area.
Normally on such a satellite there will be a lot of separate feedhorns in front of this big dish, each of them creating a separate (small) coverage area. This has two advantages: you have a lot of gain and thus a lot of signal (and not much power needed for uplink), and also the separate areas create a space-division multiplex where you can use the same frequency many times on the same satellite. Without this, it would be impossible to provide a large number of customers with a high bitrate.
Howevre, while this is all very nice and would make a good HDTV distribution satellite, the inevitably long roundtriptimes still will make it unusable for Internet access and other interactive purposes.
It does not matter anyway, because you cannot setup a small set of low-earth-orbit satellites to orbit your own country. Any satellite will always orbit the entire earth. When your coverage area is small you need just as many satellites, at the very best you could cut down on solar panel size as it could use smaller panels and charge a battery while it is not over your small country.
pinging slashdot gives me an average of 349ms from London,UK
You need to look for a better provider.... Here in the Netherlands I get:
PING slashdot.org (66.35.250.150) 56(84) bytes of data. 64 bytes from slashdot.org (66.35.250.150): icmp_seq=1 ttl=45 time=167 ms 64 bytes from slashdot.org (66.35.250.150): icmp_seq=2 ttl=45 time=167 ms 64 bytes from slashdot.org (66.35.250.150): icmp_seq=3 ttl=45 time=167 ms 64 bytes from slashdot.org (66.35.250.150): icmp_seq=4 ttl=45 time=167 ms 64 bytes from slashdot.org (66.35.250.150): icmp_seq=5 ttl=45 time=171 ms 64 bytes from slashdot.org (66.35.250.150): icmp_seq=6 ttl=45 time=167 ms
--- slashdot.org ping statistics --- 6 packets transmitted, 6 received, 0% packet loss, time 5006ms rtt min/avg/max/mdev = 167.216/168.308/171.876/1.676 ms
Even with a 33k6 modem (I use ADSL) I would not get such a long ping.
If the satellite were to be placed in a far lower orbit then latency numbers will drop. I believe this requires spin stabilisers and some sort of engine to keep the satellite from plummeting to Earth though.
The selection is between geostationary at about 36000 km, and a low-earth-orbit at say 800 km. But you would need many satellites in that orbit to have continuous coverage, as each satellite is in view for about 15 minutes and then disappears for 2 hours or more.
Projects have been proposed and even realized for that system (Motorola Iridium), but it was not financially viable and the company went bankrupt.
You are mistaken. It is completely unusable for VoIP and it seriously decreases performance of any TCP connection not specially tuned for long delay.
Some people think "but all telephony is via satellite" but this is not true at all. In the early days of international telephony, satellite links were used. But today, international calls are almost allways via cable. When you talk via satellite you notice that a lot of time the two parties are colliding with eachother because an untrained individual cannot conduct a fullduplex conversation with a second of latency.
Satellite links are only useful for unidirectional communication, like TV broadcast. Even TV newsgathering over satellite is painful to watch. Companies that setup and advertise Internet over (geostationary) satellites fall into the same category as those advocating broadband over powerlines: they have a hammer on hand and now they see everything as a nail. The power companies have their network, and see the telcos make money on Internet and they think "wanna do that too!" and presto: broadband over powerlines. Same for satellite companies. They see a market demand and propose an unsuitable solution just because they happen to be in the business.
this has not stopped a lot of popular software from being presented to the users in a distro-agnostic manner
But I think it has prevented even more software from being released for Linux! Sure you can work around it, but let's face it: for Windows you buy one of those packaging tools that handle installation on 98/2000/XP for you and you only have to supply your files and where they have to go. On Linux, it could potentially have been simpler, because the packaging tools are delivered with the OS. No need to package everything as a.EXE, you can build a.RPM that contains only your package.
But there the nightmare starts. Will your user be able to install the RPM with a commandline or via the sysadmin GUI? If the latter, where would he/she have to click? How do you write your installation manual? Does the system support RPM at all? Or do you need to supply multiple formats?
Small wonder that the few successful packages had to fall back to a custom installation program. This means a lot of extra work (proportionally more when the program to be installed is smaller, like some toolbar).
IMHO it is a major holdback for deployment of Linux on your average desktop. Users want to download something, and install it using clear instructions not longer than a few paragraphs. Like "click on the YaST icon, choose install additional software, select the program you downloaded", or "doubleclick on the program you downloaded and type the password for software installation". Not "first find out if your system uses RPM or DEB, then download the appropriate package. Or select the tarball if it supports neither of these. Then, when you chose RPM, open a Shell window and type 'rpm -Uvh the_name_of_this_program.rpm'. when you downloaded the DEB file,.... etc"
There is some internal writeup circulating about what issues were found when Microsoft converted Hotmail.com from Solaris+Linux to Windows. They had the same problems. Managing a cluster of thousands of servers was much simpeler with Linux than with Windows. Probably that has changed, but still... for this job of running a massive website cluster, Linux was more suitable than Windows. (of course they still changed it, but they had different objectives than Google)
I can understand that. Not only is Linux not yet very widely used on the desktop yet, but its userbase is also so hopelessly fragmented between "distributions" that it is quite difficult to release an end-user-installable desktop tool for Linux.
Remember that Microsoft sites have to reboot at least once every month after installing patches.
Linux sites often can avoid this (at least as far as Netcraft is concerned; restarting Apache does not cut the uptime), however there have been so many kernel updates last year that a Linux system with a year of uptime is a bit questionable as well.
(of course most kernel updates are for local exploits only; one could decide a properly firewalled system does not need them)
They claim it costs them money to handle SPAM email in their MSN Hotmail service. But when you try to make them act responsible w.r.t. spamming via their MSN members, they suddenly claim MSN is an independent company they are not responsible for...
(I cannot believe that Microsoft operates a company, MSN, that provides nearly bulletproof domain name registration an mail services for 419 scammers, and nobody at Microsoft or MSN wants to handle abuse for it. They have their abusedesk outsourced to India and the clerks there have no idea what you are talking about when you tell them the abuse@hotmail.com robot is refusing all complaints about MSN Personal Address accounts, and continue to insist that you should send the mail with full headers to that address. Which will bounce it because it is "not hotmail, not our problem")
Telephones are normally not sold directly by the providers, but by stores that sell phones and subscriptions. The providers (we have 5 physical and, I think, two virtual providers) give a premium on new subscriptions which can be had either as cash-back after some time, or as an advance towards buying a phone in the store. The premium is sufficient to completely cover the cost of a phone, and sometimes some other "free" item like a DVD recorder or TV. Attached to this, of course, is a contract of one or two years, with varying number of call minutes per month. For a prepaid card (i.e. no monthly fixed charge), the phone would cost about $50 net these days including maybe $15 worth of call minutes.
The whole market is far from clear. I would prefer if there were just phones sold for realistic prices, subscriptions without premiums, and prepaid cards. The DVD recorder or TV I can buy separately when I want one.
One would expect that on average there would be no difference it what you pay, and it would be easier to compare and select. You are right, in other countries there are strict regulations on what can be done (like handing out phones and/or premiums).
One problem is that fewer and fewer people care about quality. Even professionals realize that when they get the same money for a "digital shot that could have been better" as for a super-high-quality large negative photo (that nobody notices), they just as well might make the switch.
So while there will always be a niche for very high quality, it will become smaller and smaller all the time.
Slashdot Mirror
You are just nitpicking.
In your original post you are showing ignorance about effective radiated power by talking about "just 0.25 watts", making me try to explain that issue to the layman, and then you suddenly creep out under the cover and claim you are an extra class amateur radio operator. In that case you should have known that communication over thousands of kilometers have been made on 2.4 GHz and your claim about the power being too low is totally irrelevant.
Before turning down my calculation you should make your own math with two 2.1 dB antennas when compared with two 20 dB antennas and 0.25 W of transmitting power and you will see that with the 2.1 dB antennas you will need close to 1000 Watts of power to have the same effective signal. So what is your point?
Line of sight has never been a topic of the discussion, it was only about range and required output power. Furthermore, I don't think you can run an 11 Mbps link or even the link speed they achieved in this case on HF bands with 5 watts.
But that requires another idiot: the one that allows employees to take unprotected systems home, connect them to other networks, and then back on the company LAN.
That does not seem a very wise policy.
How did it enter your very large company?
I cannot believe that companies have port 445 open from the outside, so it must have entered a different way.
Badly firewalled laptops that pick it up while connected to the wireless internet at home and then release it back on the LAN?
You forget that they used 2ft dish antennas. An antenna like this would have about 20dB of gain. That amplifies the 0.25 watts to 25 watts!
And, it also amplifies the received signal by the same amount.
When compared with an HF bands signal transmitted by a typical dipole antenna, you would have the equivalent of about 1000 watts of power, assuming a setup like shown in the picture is used at both ends.
Of course you need line-of-sight, but the station apparently sits on a hilltop.
When comparing HF to SHF, people often don't take this factor in the calculation. In fact, when calculating the total of achievable antenna gain within a reasonable size, path loss, and achievable transmission power using off-the-shelf technology, the 2.4 GHz band is about the optimum band for getting long range line-of-sight communication. Amateur satellites have been built that transmit only a couple of watts and are (were) easily receivable over a distance of 40000 km.
Connect a decent antenna. Bundling of the radio energy is the key to covering very long distances with very small transmitted power. And it also works on the receiving end, so you can start doing this at one end and see if it helps.
A moderately directive antenna will amplify the signal by 50 times at this frequency, and is commercially available for $100 or so.
I don't get it... what is impressive about transmitting in the 2.4 GHz range?
This may have been a feat 4 decades ago, but today the parts to construct a transmitter operating in that range are commodity. Shown by the fact that anyone can buy a digital wireless transmission system for under 200 dollars.
I don't think Cisco belongs in that list. Cisco stuff is mainly overpriced hardware running an outdated monolytic piece of software that has too many bolted on additions and features. It is more or less like Windows, in that regard.
And they market it marketed so well that people still think it is the best and that it is worth the money.
But in reality the company will not even support the customer who bought their expensive hardware; they will have to pay separately for support. And small but essential feature updates cannot be made easily because everything has to go in the generally released monolythic IOS.
Cisco IOS is like running a system like Knoppix on all your production machines. You can make changes to some pre-determined configuration item but not to any part of the functionality.
Not only will it be difficult to form a public opinion on issues such as this, but also it is a widespread misbelief that public opinion has any direct effect on laws and international treaties.
Especially when patents, intellectual property and similar issues are involved, the indirect democratic system that most countries have means that the public has no influence.
You get a chance to vote a politician, yes. But people in politics to not tend to be that revolting once they are in placs like parliament.
And international business has a very effective lobbying system in place, which convinces those politicians that it is best for their countries economy to favour multinational big business.
Patents and intellectual property laws are part of that.
"voicing out loudly" is not going to make a better impression than thick reports about profits and jobs, and the occasional bribe.
Are "we" in the position to change anything about the patent system? No, we aren't.
Then what point is in discussing how to reform it, let alone carrying out those reforms?
Do you propose a coup d'etat?
At least they have discovered something:
One shortcoming of current word processing applications programs in particular is exposed when a user attempts to locate numerical information in a text document utilizing a standard "find" function.
This I noticed 20 years ago, but apparently Microsoft has now noticed it as well.
Free tip: Many Unix text editors had powerful features for searching complicated patterns even before "current word processing applications programs" were even developed. So, instead of filing yet another stupid patent, you could set a team to work and implement a decent regexp search and replace.
From where?
This usually does not help much, because on top of the TCP connection there usually exists some application protocol that sends a request, waits for an answer, and then sends the next request.
For example, if you would use such a link to provide a VPN between your home network and a network at your employer, and you would use this to logon to that network with a Windows system, you would need to wait at least 10 minutes for the logon to complete even with locally stored profile etc.
This is because Windows sends thousands of requests over the TCP connection(s) between client and server, and awaits a response on each of them before proceeding.
And then we are not even talking about thin client computing...
The current offering from Hughes, DIRECWAY, has over 250,000 broadband customers.
250,000?? In this week's news, I read that the major ADSL provider in our small country (which has 15 million inhabitants), has 1.6 million subscribers. And there are other ADSL providers (together something like 400,000 subscribers) and Cable (with over 1 million subscribers).
Internet via satellite cannot make even a small dent into that. And of course there are many reasons for that.
Antenna gain and beamwidth can be directly derived from the antenna size, when you know the frequency it operates at. For 66ft at GigaHz frequencies you will have a lot of gain and a small area.
Normally on such a satellite there will be a lot of separate feedhorns in front of this big dish, each of them creating a separate (small) coverage area. This has two advantages: you have a lot of gain and thus a lot of signal (and not much power needed for uplink), and also the separate areas create a space-division multiplex where you can use the same frequency many times on the same satellite. Without this, it would be impossible to provide a large number of customers with a high bitrate.
Howevre, while this is all very nice and would make a good HDTV distribution satellite, the inevitably long roundtriptimes still will make it unusable for Internet access and other interactive purposes.
It does not matter anyway, because you cannot setup a small set of low-earth-orbit satellites to orbit your own country.
Any satellite will always orbit the entire earth. When your coverage area is small you need just as many satellites, at the very best you could cut down on solar panel size as it could use smaller panels and charge a battery while it is not over your small country.
pinging slashdot gives me an average of 349ms from London,UK
You need to look for a better provider....
Here in the Netherlands I get:
PING slashdot.org (66.35.250.150) 56(84) bytes of data.
64 bytes from slashdot.org (66.35.250.150): icmp_seq=1 ttl=45 time=167 ms
64 bytes from slashdot.org (66.35.250.150): icmp_seq=2 ttl=45 time=167 ms
64 bytes from slashdot.org (66.35.250.150): icmp_seq=3 ttl=45 time=167 ms
64 bytes from slashdot.org (66.35.250.150): icmp_seq=4 ttl=45 time=167 ms
64 bytes from slashdot.org (66.35.250.150): icmp_seq=5 ttl=45 time=171 ms
64 bytes from slashdot.org (66.35.250.150): icmp_seq=6 ttl=45 time=167 ms
--- slashdot.org ping statistics ---
6 packets transmitted, 6 received, 0% packet loss, time 5006ms
rtt min/avg/max/mdev = 167.216/168.308/171.876/1.676 ms
Even with a 33k6 modem (I use ADSL) I would not get such a long ping.
If the satellite were to be placed in a far lower orbit then latency numbers will drop. I believe this requires spin stabilisers and some sort of engine to keep the satellite from plummeting to Earth though.
The selection is between geostationary at about 36000 km, and a low-earth-orbit at say 800 km. But you would need many satellites in that orbit to have continuous coverage, as each satellite is in view for about 15 minutes and then disappears for 2 hours or more.
Projects have been proposed and even realized for that system (Motorola Iridium), but it was not financially viable and the company went bankrupt.
You are mistaken. It is completely unusable for VoIP and it seriously decreases performance of any TCP connection not specially tuned for long delay.
Some people think "but all telephony is via satellite" but this is not true at all. In the early days of international telephony, satellite links were used. But today, international calls are almost allways via cable.
When you talk via satellite you notice that a lot of time the two parties are colliding with eachother because an untrained individual cannot conduct a fullduplex conversation with a second of latency.
Satellite links are only useful for unidirectional communication, like TV broadcast. Even TV newsgathering over satellite is painful to watch.
Companies that setup and advertise Internet over (geostationary) satellites fall into the same category as those advocating broadband over powerlines: they have a hammer on hand and now they see everything as a nail.
The power companies have their network, and see the telcos make money on Internet and they think "wanna do that too!" and presto: broadband over powerlines.
Same for satellite companies. They see a market demand and propose an unsuitable solution just because they happen to be in the business.
Don't fall for it!
this has not stopped a lot of popular software from being presented to the users in a distro-agnostic manner
.EXE, you can build a .RPM that contains only your package.
.... etc"
But I think it has prevented even more software from being released for Linux!
Sure you can work around it, but let's face it: for Windows you buy one of those packaging tools that handle installation on 98/2000/XP for you and you only have to supply your files and where they have to go.
On Linux, it could potentially have been simpler, because the packaging tools are delivered with the OS. No need to package everything as a
But there the nightmare starts. Will your user be able to install the RPM with a commandline or via the sysadmin GUI? If the latter, where would he/she have to click? How do you write your installation manual?
Does the system support RPM at all? Or do you need to supply multiple formats?
Small wonder that the few successful packages had to fall back to a custom installation program. This means a lot of extra work (proportionally more when the program to be installed is smaller, like some toolbar).
IMHO it is a major holdback for deployment of Linux on your average desktop. Users want to download something, and install it using clear instructions not longer than a few paragraphs. Like "click on the YaST icon, choose install additional software, select the program you downloaded", or "doubleclick on the program you downloaded and type the password for software installation".
Not "first find out if your system uses RPM or DEB, then download the appropriate package. Or select the tarball if it supports neither of these. Then, when you chose RPM, open a Shell window and type 'rpm -Uvh the_name_of_this_program.rpm'. when you downloaded the DEB file,
There is some internal writeup circulating about what issues were found when Microsoft converted Hotmail.com from Solaris+Linux to Windows.
They had the same problems. Managing a cluster of thousands of servers was much simpeler with Linux than with Windows. Probably that has changed, but still... for this job of running a massive website cluster, Linux was more suitable than Windows.
(of course they still changed it, but they had different objectives than Google)
I can understand that. Not only is Linux not yet very widely used on the desktop yet, but its userbase is also so hopelessly fragmented between "distributions" that it is quite difficult to release an end-user-installable desktop tool for Linux.
Remember that Microsoft sites have to reboot at least once every month after installing patches.
Linux sites often can avoid this (at least as far as Netcraft is concerned; restarting Apache does not cut the uptime), however there have been so many kernel updates last year that a Linux system with a year of uptime is a bit questionable as well.
(of course most kernel updates are for local exploits only; one could decide a properly firewalled system does not need them)
If SCO put Linux GPL code in their proprietary product, violating the GPL, the evidence can be found in 30 seconds with grep
Positive evidence, yes.
But negative evidence (prove that there is no such code) is a little harder...
They claim it costs them money to handle SPAM email in their MSN Hotmail service.
But when you try to make them act responsible w.r.t. spamming via their MSN members, they suddenly claim MSN is an independent company they are not responsible for...
(I cannot believe that Microsoft operates a company, MSN, that provides nearly bulletproof domain name registration an mail services for 419 scammers, and nobody at Microsoft or MSN wants to handle abuse for it. They have their abusedesk outsourced to India and the clerks there have no idea what you are talking about when you tell them the abuse@hotmail.com robot is refusing all complaints about MSN Personal Address accounts, and continue to insist that you should send the mail with full headers to that address. Which will bounce it because it is "not hotmail, not our problem")
Telephones are normally not sold directly by the providers, but by stores that sell phones and subscriptions.
The providers (we have 5 physical and, I think, two virtual providers) give a premium on new subscriptions which can be had either as cash-back after some time, or as an advance towards buying a phone in the store.
The premium is sufficient to completely cover the cost of a phone, and sometimes some other "free" item like a DVD recorder or TV.
Attached to this, of course, is a contract of one or two years, with varying number of call minutes per month.
For a prepaid card (i.e. no monthly fixed charge), the phone would cost about $50 net these days including maybe $15 worth of call minutes.
The whole market is far from clear. I would prefer if there were just phones sold for realistic prices, subscriptions without premiums, and prepaid cards. The DVD recorder or TV I can buy separately when I want one.
One would expect that on average there would be no difference it what you pay, and it would be easier to compare and select.
You are right, in other countries there are strict regulations on what can be done (like handing out phones and/or premiums).
One problem is that fewer and fewer people care about quality.
Even professionals realize that when they get the same money for a "digital shot that could have been better" as for a super-high-quality large negative photo (that nobody notices), they just as well might make the switch.
So while there will always be a niche for very high quality, it will become smaller and smaller all the time.