If you're talking about NTSC over twisted-pair, I'd recommend against it. I notice significant rollof of high frequencies with high-quality baluns on Cat 6. Cat 5 is much worse. Coax is still the cable of choice for video.
Actually, I'm talking about PAL over Cat5, because I'm in Europe;-). I agree baluns are not the way to do it - however, there exist active solutions which can be set up to compensate for the HF rolloff. Reviews that I've read on such units (eg, google for Kat5), suggest that over normal domestic distances, they're fine.
What about running Svideo and RCA to everyroom or an extra drop of cat5 to run sound on?
There are various (cheapish - under $100) ways of running video over Cat5. Personally, I've got Cat5 and audio cables into rooms, and run everything over that.
However, 2 per room is not enough. There are a number of reasons for this:
They're never where you want them. You'll always want to move stuff around, and cables draped across the floor doesn't look good.
Once you've got the cables in, you'll see more uses. For example, telephone, TV, Audio for hifi, remote control for hifi, remote control for lighting, PC network - all possibly in one room.
When building, it's cheap. It's much easier to put these cables in before the drywall goes up. Trying to fish them through afterwards is hard work. BTDT
Don't bother with fibre. It's not going to get used. People have been saying "fibre is the next big thing" for 5 years now - and cable just keeps catching up. Video over Cat5 is cheap - so, instead of piping cable round the house, put all the set top boxes in one location, and remote control them. 100Mb Ethernet is very cheap - VHS video quality needs about 3 Mb/sec, so you can stream that across existing networks. Hi definition uncompressed video is more than 100Mb/sec - but 1Gb/sec over copper is on the horizon.
Game ports don't actually have ADC's on them. The standard design is to discharge a capacitor, then charge it through the resitance of the joystick pot, and tim ehow long it takes to get to a specific voltage. This is related to the resistance. So, you can't use the game port for measuring voltages directly.
Picotech have addons for PCs to convert into a datalogging scope. I've never used one - and note that they tend to connect via the Parallel port. You can't get a whole lot of data through that port, so a 100Mhz scope won't actually show on the screen in real time. This doesn't usually matter, but you do want to bear in mind the delay between the measurement being made, and it actually appearing on the PC monitor.
On the plus side, they normally have data processing facilities that you only find on top of the range digital scopes - things like "trigger before", and FFT's.
I'm not sure why you want to go multihomed, with all the attendant problems that it brings. If this is a corporate connection, that's not got services (other than mail) being provided to the outside world, then I don't really see the point. I think you can provide the redundancy in other ways - here are some ideas, using 2 ISPs (and PA IP addresses allocated by each of them).
Put a mail server on each connection (or map an IP address from each connection through your firewall to the mail server). MX records will do your load balancing and redundancy for you.
Use NAT/PAT for users to connect to the Internet. If one conenction goes down, remove the internal routing to that connection - all your sessions will now go out of the other connection. I find that this is quicker than waiting for BGP to reroute connections via a backup/alternate path. It also gives you more flexibility in internal network numbering, and to move ISPs.
Host services with colocation providers - not internally. Colo service providers have already solved most of the service provision problems, and are well connected to the Internet - I don't think it's worth trying to do this in house.
You haven't explained why you want to use a remote machine.
If it's to run applications, then the WinVNC/PCAnywhere/etc route is a good one. It's more flexible, because you don't need a cable. You can share the machine between more than one person, and you don't need extra hardware.
Personally, for remote administration, I'd always use the KVM extender solution. If the machine that you're administrating isn't behaving, then the remote control software probably isn't working properly either. Remote software doesn't let you watch bootup screens, or reconfigure the Bios.
So, before you make a decision, I think you need to look carefully at the purpose of the solution.
One of the goals listed in the article was to allow people to communicate amongst themselves, and to play games etc across the network. I don't think there's actually a need to offer Internet connection across such a network - in which case, the costs just become the cost of the access point.
If you're not connected to the Internet, and you're not selling service, I'm not sure (I'm not in the USA, so I also don't care:) ) if you would be viewed as providing a network service. It's very similar to the Amateur Radio digipeater/node network, so I'd guess not.
So, the question is, do people want to use this network to talk amongst themselves, or to share the price of an Internet connection?
If you have SO MUCH MUSIC that you need to rip to mp3s so that you can easily get to it all why not get a 100 (150 anyone?) CD changer?
I have a 200 CD player with my CD's in it. However, only one person can listen to it at a time - so I can't listen to one CD and womone else in the house listen to another from the same player. I'm looking for a good solution to that, and this might be it.
I was probably having a bad day when I tried it - I just remember spending a morning trying to get it to create tables and allow remote access, and couldn't. I may well have been trying it on a Slackware box, rather than RedHat - I certainly remember building it from source at the time.
I can't comment on the administration aspects of PostgreSQL, because I've never run it. The point I was trying to make (several comments up in this thread) is that there are a number of situations where you don't need the facilities that Oracle/ SQL Server/ PostgreSQL/ DB2/ etc give you, and for those, MySQL is a perfectly good choice.
I know this may seem like a radical idea, but did you consider using your file systems for this?
Yes. However, the key part is distributed. The database sits behind a bunch of web servers, which get content from the database. If the primary database fails, I can very easily change to the backup - no synchronisation issues like I might have with a distributed file system. File systems have issues with lots of files in a single directory, so I'd have had to implement some sort of hierarchial file store.
Forcing the writes from one client to the server in time for another client to pick them up is another issue with the particular application I'm using.
I've been looking for a while at distributed file systems such as Coda or NFS to solve some of these problems, but they either don't seem ready for production, or seem overkill for what I'm doing. I use NFS internally for sharing files, but I don't trust it for production web servers - the applications can't tell quickly that the file system is toast, and they need to switch to the backup.
Why do you imply that postgresql is hard to setup?
It's very easy, a piece of cake.
Maybe I worded that wrong: I didn't mean to imply that postgresql is hard to setup - I meant that I found it difficult to install, when I was making the database decision a couple of years ago. I spent a morning trying to get it to work, and failed - which reflects much more on my (lack of) knowledge of PostgreSQL than the product itself.
So you don't
need the features of postgresql. Does it hurt to
have them on hand if you need them in the future?
No, unless they get in the way - which in this case they did. This particular application is very self contained - I'm never going to need the other features. When I process the data, I move it into another database (Oracle) and then I've got all the features I need - such as materialized views, transportable tablespaces, cube queries, bitmapped indexes.
I use MySQL for one of my web based applications, and I use Oracle for others. For the one I use MySQL for, I chose it because:
- It's easy to run.
I don't have to do much (if any) maintenance and management - I don't need to check if redo logs are too small, don't need to check for extents growing out of control etc.
- It's fast.
For my application - with simple inserts and deletes, MySQL is really quick. That saves me money - I can get away with a single processor Linux box for my database server, rather than something much more expensive.
- I don't need the features of a larger database.
I'm using MySQL to store and retrive information - basically a distributed file system. I don't need clever locking, transactions, views, foreign keys, triggers, stored procedures. On the applications I do need those features, I use Oracle.
- It's cheap
I looked at PostgreSQL, but it was going to take me a while to figure out how to get it set up - I find MySQL very easy to install and get running. Oracle was going to be very expensive - and it's not that easy to install and get running properly.
When you add in the costs of installation, learning how the software behaves, and the management time in keeping the software running, MySQL (for me) came out as the most cost effective option.
Oh, yeah, the TV has a circut for decryption, couldn't I just grab the digital signal on the other side of the decryption chip?
The output of that chip is probably suitable for driving the tube - so it's RGB signals for the guns, and sync for the coils. So, no, you can't just grab that signal, without some extra processing. At least, that's the way I'd do it if I was designing the stuff.
I don't think the "I can't record" is going to be an issue. The media companies seem to want to move to a pay-per-view model - so you can watch anything whenever you want, as long as you pay every time. Also - (at least in the UK) the media companies can selectively decide what they don't want you to record on digital systems - so they could easily let you record soaps etc, but not the big films.
It's happened. All Sky Digiboxes have to have Macrovision capable outputs, and macrovision is enabled over the air for certain programs. It's even in the contracts that they won't supply certain programmes if your digibox can't do macrovision. The only way you can watch Sky Digital is with a Sky approved digibox, which must run Sky approved software, and come supplied with the Sky remote control (so that they can say "press the red button").
Digital input TVs are now appearing in the UK, that take the satellite signal directly - so the decrypted programming doesn't appear outside the TV. I expect these to grow in popularity, and have TiVo like functionality soon.
As for handshaking with the broadcaster (mentioned elsewhere), if you don't have your Digibox connected to a telephone line, you have to pay upwards of £300 ($420) for the box. If you connect it, the box is free. So, in the UK, almost all of the boxes are connected to a phone line - and they phone home at intervals.
If you're running something like a 75xx series router, there are multiple processors. To cope with high load, you run VIP cards. Each VIP card accepts two interface cards, which have, for example 4 serial ports.
The interface connects to the VIP over a dedicated PCI bus - so each interface has it's own PCI to teh VIP. The VIP has its own processor on board - if you run distributed switching, then the VIP has the full forwarding table, and can route packets to other interfaces. So, in any high end router, you've got multiple processors all talking to each other, with the master processor talking to other routers and keeping the routing tables updated.
If you know the commands, you can actually log in to the VIP cards - they run a full IOS.
Separately shipped for a nominal fee covering copying costs to anybody so wishing to receive them, irreguardless of their purchase of the original product
Not sure about this - where does it say I have to ship source to non purchasers? As far as I know, the GPL doesn't say anything about people I haven't sold to - there is nothing stopping me refusing to sell my binaries to anyone - I just can't stop any of my customers doing that.
It's been done. I worked on a product to limit bandwidth to certain routes more than others - specifically, you could buy say 512k European access, and 256k World access, and 128k national access - all down the same link. No magic - just work around a few Cisco bugs, and it works.
Aha! That's what the monthly subscription charge is for.
There's a key point here - the price is not related to the cost to the supplier. The pricing structure is designed to allow the supplier to make a profit. For example, it's currently free to reconnect a telephone to BT if there has been service in the past at an address. Cost to BT: sending an engineer out to repatch, recommission line. Advantage to BT: new customer. The actual marginal cost to the supplier of the telephone call is almost zero (just the extra cost of electricity to drive a telephone in use) for local or national calls, assuming they own the cables etc.
Usage based billing is just a different way of distributing the costs.
Per minute for all calls. There are 3 "Inland" bands - local, regional (up to 35 miles) and long distance, plus international/ 1-900 type/etc. Local calls are 1 pence/minute weekends, 1.5 pence/minute evenings, and 3.95 pence/minute during the day. Full details from BT's website: http://www.s erviceview.bt.com/list/current/docs/Call_Charges/0 0161.htm
British Interactive Broadcasting "give" you a £167 subsidy, on condition that you connect the box to a telephone line. This connection is checked (I don't know how often - I don't have a box). If you disconnect from the phone line, (or, more correctly, if BIB detect that your box has not dialled them for some undisclosed period of time), you get a bill for the subsidy. So, it's not a fine, but it's as good as. You can buy a box without the subsidy, but I understand that it has been "suggested" to many larger retailers to not bring this to the attention of the buyer as strongly as they may otherwise wish to....
Slashdot Mirror
Actually, I'm talking about PAL over Cat5, because I'm in Europe ;-). I agree baluns are not the way to do it - however, there exist active solutions which can be set up to compensate for the HF rolloff. Reviews that I've read on such units (eg, google for Kat5), suggest that over normal domestic distances, they're fine.
There are various (cheapish - under $100) ways of running video over Cat5. Personally, I've got Cat5 and audio cables into rooms, and run everything over that.
However, 2 per room is not enough. There are a number of reasons for this:
Don't bother with fibre. It's not going to get used. People have been saying "fibre is the next big thing" for 5 years now - and cable just keeps catching up. Video over Cat5 is cheap - so, instead of piping cable round the house, put all the set top boxes in one location, and remote control them. 100Mb Ethernet is very cheap - VHS video quality needs about 3 Mb/sec, so you can stream that across existing networks. Hi definition uncompressed video is more than 100Mb/sec - but 1Gb/sec over copper is on the horizon.
Game ports don't actually have ADC's on them. The standard design is to discharge a capacitor, then charge it through the resitance of the joystick pot, and tim ehow long it takes to get to a specific voltage. This is related to the resistance. So, you can't use the game port for measuring voltages directly.
Picotech have addons for PCs to convert into a datalogging scope. I've never used one - and note that they tend to connect via the Parallel port. You can't get a whole lot of data through that port, so a 100Mhz scope won't actually show on the screen in real time. This doesn't usually matter, but you do want to bear in mind the delay between the measurement being made, and it actually appearing on the PC monitor.
On the plus side, they normally have data processing facilities that you only find on top of the range digital scopes - things like "trigger before", and FFT's.
I'm not sure why you want to go multihomed, with all the attendant problems that it brings. If this is a corporate connection, that's not got services (other than mail) being provided to the outside world, then I don't really see the point. I think you can provide the redundancy in other ways - here are some ideas, using 2 ISPs (and PA IP addresses allocated by each of them).
Put a mail server on each connection (or map an IP address from each connection through your firewall to the mail server). MX records will do your load balancing and redundancy for you.
Use NAT/PAT for users to connect to the Internet. If one conenction goes down, remove the internal routing to that connection - all your sessions will now go out of the other connection. I find that this is quicker than waiting for BGP to reroute connections via a backup/alternate path. It also gives you more flexibility in internal network numbering, and to move ISPs.
Host services with colocation providers - not internally. Colo service providers have already solved most of the service provision problems, and are well connected to the Internet - I don't think it's worth trying to do this in house.
You haven't explained why you want to use a remote machine.
If it's to run applications, then the WinVNC/PCAnywhere/etc route is a good one. It's more flexible, because you don't need a cable. You can share the machine between more than one person, and you don't need extra hardware.
Personally, for remote administration, I'd always use the KVM extender solution. If the machine that you're administrating isn't behaving, then the remote control software probably isn't working properly either. Remote software doesn't let you watch bootup screens, or reconfigure the Bios.
So, before you make a decision, I think you need to look carefully at the purpose of the solution.
One of the goals listed in the article was to allow people to communicate amongst themselves, and to play games etc across the network. I don't think there's actually a need to offer Internet connection across such a network - in which case, the costs just become the cost of the access point.
If you're not connected to the Internet, and you're not selling service, I'm not sure (I'm not in the USA, so I also don't care :) ) if you would be viewed as providing a network service. It's very similar to the Amateur Radio digipeater/node network, so I'd guess not.
So, the question is, do people want to use this network to talk amongst themselves, or to share the price of an Internet connection?
I have a 200 CD player with my CD's in it. However, only one person can listen to it at a time - so I can't listen to one CD and womone else in the house listen to another from the same player. I'm looking for a good solution to that, and this might be it.
I was probably having a bad day when I tried it - I just remember spending a morning trying to get it to create tables and allow remote access, and couldn't. I may well have been trying it on a Slackware box, rather than RedHat - I certainly remember building it from source at the time.
I can't comment on the administration aspects of PostgreSQL, because I've never run it. The point I was trying to make (several comments up in this thread) is that there are a number of situations where you don't need the facilities that Oracle/ SQL Server/ PostgreSQL/ DB2/ etc give you, and for those, MySQL is a perfectly good choice.
Yes. However, the key part is distributed. The database sits behind a bunch of web servers, which get content from the database. If the primary database fails, I can very easily change to the backup - no synchronisation issues like I might have with a distributed file system. File systems have issues with lots of files in a single directory, so I'd have had to implement some sort of hierarchial file store.
Forcing the writes from one client to the server in time for another client to pick them up is another issue with the particular application I'm using.
I've been looking for a while at distributed file systems such as Coda or NFS to solve some of these problems, but they either don't seem ready for production, or seem overkill for what I'm doing. I use NFS internally for sharing files, but I don't trust it for production web servers - the applications can't tell quickly that the file system is toast, and they need to switch to the backup.
I use MySQL for one of my web based applications, and I use Oracle for others. For the one I use MySQL for, I chose it because:
- It's easy to run.
I don't have to do much (if any) maintenance and management - I don't need to check if redo logs are too small, don't need to check for extents growing out of control etc.
- It's fast.
For my application - with simple inserts and deletes, MySQL is really quick. That saves me money - I can get away with a single processor Linux box for my database server, rather than something much more expensive.
- I don't need the features of a larger database.
I'm using MySQL to store and retrive information - basically a distributed file system. I don't need clever locking, transactions, views, foreign keys, triggers, stored procedures. On the applications I do need those features, I use Oracle.
- It's cheap
I looked at PostgreSQL, but it was going to take me a while to figure out how to get it set up - I find MySQL very easy to install and get running. Oracle was going to be very expensive - and it's not that easy to install and get running properly.
When you add in the costs of installation, learning how the software behaves, and the management time in keeping the software running, MySQL (for me) came out as the most cost effective option.
Colin.
The output of that chip is probably suitable for driving the tube - so it's RGB signals for the guns, and sync for the coils. So, no, you can't just grab that signal, without some extra processing. At least, that's the way I'd do it if I was designing the stuff.
I don't think the "I can't record" is going to be an issue. The media companies seem to want to move to a pay-per-view model - so you can watch anything whenever you want, as long as you pay every time. Also - (at least in the UK) the media companies can selectively decide what they don't want you to record on digital systems - so they could easily let you record soaps etc, but not the big films.
It's happened. All Sky Digiboxes have to have Macrovision capable outputs, and macrovision is enabled over the air for certain programs. It's even in the contracts that they won't supply certain programmes if your digibox can't do macrovision. The only way you can watch Sky Digital is with a Sky approved digibox, which must run Sky approved software, and come supplied with the Sky remote control (so that they can say "press the red button").
Digital input TVs are now appearing in the UK, that take the satellite signal directly - so the decrypted programming doesn't appear outside the TV. I expect these to grow in popularity, and have TiVo like functionality soon.
As for handshaking with the broadcaster (mentioned elsewhere), if you don't have your Digibox connected to a telephone line, you have to pay upwards of £300 ($420) for the box. If you connect it, the box is free. So, in the UK, almost all of the boxes are connected to a phone line - and they phone home at intervals.
If you're running something like a 75xx series router, there are multiple processors. To cope with high load, you run VIP cards. Each VIP card accepts two interface cards, which have, for example 4 serial ports.
The interface connects to the VIP over a dedicated PCI bus - so each interface has it's own PCI to teh VIP. The VIP has its own processor on board - if you run distributed switching, then the VIP has the full forwarding table, and can route packets to other interfaces. So, in any high end router, you've got multiple processors all talking to each other, with the master processor talking to other routers and keeping the routing tables updated.
If you know the commands, you can actually log in to the VIP cards - they run a full IOS.
Already done for a number of years. Try installing SunSoft software under Solaris on a Sun box, or HP OpenView.
Colin.
Separately shipped for a nominal fee covering copying costs to anybody so wishing to receive them, irreguardless of their purchase of the original product
Not sure about this - where does it say I have to ship source to non purchasers? As far as I know, the GPL doesn't say anything about people I haven't sold to - there is nothing stopping me refusing to sell my binaries to anyone - I just can't stop any of my customers doing that.
It's been done. I worked on a product to limit bandwidth to certain routes more than others - specifically, you could buy say 512k European access, and 256k World access, and 128k national access - all down the same link. No magic - just work around a few Cisco bugs, and it works.
Couple of key clarifications:
Demon is _not_ Dr. Godfreys ISP.
Other ISPs were also contacted, and AIUI refused to delete the article. Dr. Godfrey has only initiated legal proceedings (AFAIK) against Demon.
The court case is ongoing - Demon have not lost. They have lost some pretrial bits - but I'm not sure what.
Cheers - Colin.
Aha! That's what the monthly subscription charge is for.
There's a key point here - the price is not related to the cost to the supplier. The pricing structure is designed to allow the supplier to make a profit. For example, it's currently free to reconnect a telephone to BT if there has been service in the past at an address. Cost to BT: sending an engineer out to repatch, recommission line. Advantage to BT: new customer. The actual marginal cost to the supplier of the telephone call is almost zero (just the extra cost of electricity to drive a telephone in use) for local or national calls, assuming they own the cables etc.
Usage based billing is just a different way of distributing the costs.
Per minute for all calls. There are 3 "Inland" bands - local, regional (up to 35 miles) and long distance, plus international/ 1-900 type/etc. Local calls are 1 pence/minute weekends, 1.5 pence/minute evenings, and 3.95 pence/minute during the day. Full details from BT's website: http://www.s erviceview.bt.com/list/current/docs/Call_Charges/0 0161.htm
British Interactive Broadcasting "give" you a £167 subsidy, on condition that you connect the box to a telephone line. This connection is checked (I don't know how often - I don't have a box). If you disconnect from the phone line, (or, more correctly, if BIB detect that your box has not dialled them for some undisclosed period of time), you get a bill for the subsidy. So, it's not a fine, but it's as good as. You can buy a box without the subsidy, but I understand that it has been "suggested" to many larger retailers to not bring this to the attention of the buyer as strongly as they may otherwise wish to....