Notices
Results 1 to 19 of 19

Thread: Ethernet and Bridges

  1. #1 Ethernet and Bridges 
    Forum Freshman
    Join Date
    Feb 2009
    Posts
    7
    Can an Ethernet be extended to arbitrary lengths using bridges? I want to know if the answer is an yes or a no?

    HM


    Reply With Quote  
     

  2.  
     

  3. #2 Re: Ethernet and Bridges 
    Suspended
    Join Date
    Apr 2008
    Posts
    2,176
    Quote Originally Posted by R0jkumar
    Can an Ethernet be extended to arbitrary lengths using bridges? I want to know if the answer is an yes or a no?

    HM
    Not sure exactly what you mean by bridge, my terminology is a bit rusty. However you can through a server or router connect just about any Ethernet, either, buss-star or buss ring, buss loop, twisted pair or any other system, to any other system. Even old mainframe. And I am pretty sure you can connect the new mainframes from IBM today. Ha-ha.

    Originally there was STAR, then there was, this cross between a buss loop and star. It was pretty silly. But the name brand Ethernet was actually if I am not mistaken used on a couple different systems and the Network cards for those systems.

    I actually used Etherlink cards by Western Digital in the eighties to create a 10 mebabyte a second buss loop. It was sold as 10 mega "byte" a second buss loop. Using the 50 ohm BNC connectors. It screamed for the time. The cards were better then the computers at the time. It was stated to support 70 users.

    The way I found out how fast the Network was, was the day, I got faster computers. I could not believe how fast it downloaded the hard drive, from the server. Faster then a local hard drive could absorb it.

    A fellow who did the computers for Avis, ordered the equipment for me, he was very good. He repaired all the hard to fix stuff at Avis.

    The network passed data at the speed of 10 mega bytes a second to the computers, but the computers, just had nothing to do with the data.


    Sometimes I find myself in server farms. And I was talking to the guys that program them, and all the different types of networks and backup networks. And they were saying that, they can make any data appear to be anywhere. By the use of servers/routers connecting different networks.

    You can connect one network to another. If you give your trust to some other network. But you cannot trust any other network really unless you trust everyone in that other network. Because when someone in the other network connects to another network, your information could end up anywhere.

    In fact today just because you use a server farm here in the United States, and if you do a visual trace, using McAfee Virus scan, tools. You can see where another site is hosted from. But you can only see where they want you to see the other site is hosted from. Your stuff could be in Bangladesh. And even a tracing program would not know it.


    Here is a tracing program in action.

    http://www.Rockwelder.com/Flash/Tracer/Tracer.html



    Sincerely,


    William McCormick


    Reply With Quote  
     

  4. #3  
    Forum Masters Degree Numsgil's Avatar
    Join Date
    Jan 2009
    Posts
    708
    Arbitrary length, probably not. There's signal degradation in Cat5 over long distances, and I don't think even using repeaters that you're going to overcome that. Like I don't think you could possibly run an underground Cat5 cable with repeaters from Cal Tek to Harvard. But for any reasonable distance, like inside an office building, I think the answer is probably yes.

    Or did you mean arbitrary size, like number of computers on the network?
    Reply With Quote  
     

  5. #4 Ethernet and Bridges 
    Forum Freshman
    Join Date
    Feb 2009
    Posts
    7
    Here is my concern. On one hand there is some literature suggestion that an ethernet using switches can be arbitrarily long.

    In Section 5.6.4, Page 467 of the book "Computer Networking - Top Down Approach " it is mentioned that in theory , using switches , it is possible to build a LAN that spans the entire globe. It is also mentioned on page 471 that when switches are used as interconnection devices, there is no theoretical limit to the geographical reach of a LAN.

    But i have come across some literature including the wikipedia website (http://en.wikipedia.org/wiki/Network_bridge ) which mention that one of the disadvantages of network bridges is that they do not scale to extremely large networks. But then it goes on to mention that

    "Although infinite bridges(or layer 2 switches) can be connected in theory, often a broadcast storm will result as more and more collisions occur".

    I do know that an ethernet using just repeaters cannot be more than a fixed length long since the nodes in such an ethernet would be in the same collision detection domain and they would not be able to detect collisions. When i say length, i mean the physical length and hence by extension the number of nodes.

    Hence, if there is a yes or no question of the form "Can an ethernet be extended to arbitrary length using bridges"? , would an answer be a yes or a no?

    RK
    Reply With Quote  
     

  6. #5 Re: Ethernet and Bridges 
    Suspended
    Join Date
    Apr 2008
    Posts
    2,176
    Quote Originally Posted by R0jkumar
    Here is my concern. On one hand there is some literature suggestion that an ethernet using switches can be arbitrarily long.

    In Section 5.6.4, Page 467 of the book "Computer Networking - Top Down Approach " it is mentioned that in theory , using switches , it is possible to build a LAN that spans the entire globe. It is also mentioned on page 471 that when switches are used as interconnection devices, there is no theoretical limit to the geographical reach of a LAN.

    But i have come across some literature including the wikipedia website (http://en.wikipedia.org/wiki/Network_bridge ) which mention that one of the disadvantages of network bridges is that they do not scale to extremely large networks. But then it goes on to mention that

    "Although infinite bridges(or layer 2 switches) can be connected in theory, often a broadcast storm will result as more and more collisions occur".

    I do know that an ethernet using just repeaters cannot be more than a fixed length long since the nodes in such an ethernet would be in the same collision detection domain and they would not be able to detect collisions. When i say length, i mean the physical length and hence by extension the number of nodes.

    Hence, if there is a yes or no question of the form "Can an ethernet be extended to arbitrary length using bridges"? , would an answer be a yes or a no?

    RK
    A buss loop system, uses a method, of sending packets out, based on time. That is why you can only have about 70 users to a loop.
    Each computer added to a loop, causes a time delay in the servers, that are logged onto the Network or servers.
    Each server has to listen for known computers logged onto the loop, in the time window allotted to them, when they log onto the network/server.
    To connect one network to another, you have to basically connect one loop to another and another, through a server or gateway/router.

    Each device added to a loop, adds a time window of about 1/70th of the loops total capability. So you can only add about 69 new loops to a loop, and then you can only have one user able to access that single terminal in the original loop. But he has access to 69 loops, and their sub loops.

    There are problems doing all this obviously.

    There are different ways a server logs onto a loop and a terminal logs onto a loop. Each server logged on requires more time, to allow for it to go through the packets addressed to it, from all the computers in the loop.

    They over come this with special equipment at local communications switching stations. Where all the local loops congregate. Than they route them to other switching communications centers, far away. Very much like old telephone communications.



    Sincerely,


    William McCormick
    Reply With Quote  
     

  7. #6  
    Suspended
    Join Date
    Apr 2008
    Posts
    2,176
    You can use modems to connect one router to another. So one router/modem on a loop, can be connected to another loop far away by another router or modem.

    It is hard to explain, because, although anything is possible, I mean anything. A lot of the equipment is only designed to do certain things.

    However once you start writing your own stuff. Forget it, you can fake anything communicate anything, anywhere. A single terminal connected to a network can communicate just about anything in that network to anywhere.

    Some farms use Satellite as a backup, incase the heavy hard wire crashes, with a power outage. They can just switch over their whole system to the Satellite system.

    That is why when companies were claiming so much security, and they were letting people download stuff. I thought wow, they are really experts, Ha-ha.

    If you did not write the bios loop, the operating system, the software, the network communications systems. They own you. If you download something they own you.

    So the answer was Yes. You can send your network loop anywhere.



    Sincerely,


    William McCormick
    Reply With Quote  
     

  8. #7  
    Forum Isotope (In)Sanity's Avatar
    Join Date
    Oct 2004
    Location
    Mesa AZ
    Posts
    2,699
    A bridge is just a repeater of packets, so yes it can be extended to any length desired. It could span the entire universe if you wanted. The TCP/IP stack would have to be modified a bit to account for the many years of latency between packets however

    Most switches are just mufti port bridges with intelligent routing based on MAC addresses.
    Pleased to meet you. Hope you guess my name
    Reply With Quote  
     

  9. #8  
    Forum Masters Degree Numsgil's Avatar
    Join Date
    Jan 2009
    Posts
    708
    The problem I foresee if you're spanning hundreds of miles is packet loss. The signal ends up mutating a bit here and there. Sometimes its correctable, sometimes not. At extremely long distances the tiny mutations add up to an increasing packet loss. So as you approach thousands of miles the packet loss is probably like 99%.

    That's just my guess. I'm not a hardware guy :P
    Reply With Quote  
     

  10. #9  
    Forum Isotope (In)Sanity's Avatar
    Join Date
    Oct 2004
    Location
    Mesa AZ
    Posts
    2,699
    Quote Originally Posted by Numsgil
    The problem I foresee if you're spanning hundreds of miles is packet loss. The signal ends up mutating a bit here and there. Sometimes its correctable, sometimes not. At extremely long distances the tiny mutations add up to an increasing packet loss. So as you approach thousands of miles the packet loss is probably like 99%.

    That's just my guess. I'm not a hardware guy :P
    Sorry, but the internet has proven otherwise. Most of our posters are hitting this server from thousands of miles away. Your correct in theory, in practice the percentage is really very low. Also the bridge is just acting like a repeater, it's up to the client server combination and the TCP/IP stack to determine what packets were lost and what to resend.
    Pleased to meet you. Hope you guess my name
    Reply With Quote  
     

  11. #10  
    Forum Masters Degree Numsgil's Avatar
    Join Date
    Jan 2009
    Posts
    708
    Most of the internet travels along either fiber optics or satellites (satellites are why there's several ms of latency in online games. Speed of light lag getting from ground to geosynch orbit and back). So that's not a relevant example.
    Reply With Quote  
     

  12. #11  
    Forum Isotope (In)Sanity's Avatar
    Join Date
    Oct 2004
    Location
    Mesa AZ
    Posts
    2,699
    Quote Originally Posted by Numsgil
    Most of the internet travels along either fiber optics or satellites (satellites are why there's several ms of latency in online games. Speed of light lag getting from ground to geosynch orbit and back). So that's not a relevant example.
    What speed do you think signals travel over CAT5 ? Now that you looked that up is it faster , slower or the same speed as the microwaves used for satellite communications? The majority of people are not using satellites for their internet access, even on mid hops. The use of optical is a given, I still am not sure why you would not see my example as relevant to the question ?
    Pleased to meet you. Hope you guess my name
    Reply With Quote  
     

  13. #12  
    Forum Professor leohopkins's Avatar
    Join Date
    Dec 2006
    Location
    Dulwich, London, England
    Posts
    1,418
    Okay, just to tidy a few things up and answer a few questions within this thread......

    Ethernet was devised by Xerox.

    Also, On a normal run, cat5 100BaseT cable will support lengths of 98 metres before signal loss, after that repeaters must be installed. Both switches and bridges and routers also perform repeating functions so these could be used to extend the maximum cable length to the moon and back if you so wish. The only thing to remember here is that the more components (i.e in this case intermediary devices) you have, the more things can go wrong when a device malfunctions.

    As for collisions happening more frequently, that would depend entirely on the rest of your network topology. If you have a 6km ethernet backbone propped up with bidges to repeat the signal in one big collission domain then, yes things will go wrong.
    The hand of time rested on the half-hour mark, and all along that old front line of the English there came a whistling and a crying. The men of the first wave climbed up the parapets, in tumult, darkness, and the presence of death, and having done with all pleasant things, advanced across No Man's Land to begin the Battle of the Somme. - Poet John Masefield.

    www.leohopkins.com
    Reply With Quote  
     

  14. #13  
    Forum Masters Degree Numsgil's Avatar
    Join Date
    Jan 2009
    Posts
    708
    Quote Originally Posted by (In)Sanity
    Quote Originally Posted by Numsgil
    Most of the internet travels along either fiber optics or satellites (satellites are why there's several ms of latency in online games. Speed of light lag getting from ground to geosynch orbit and back). So that's not a relevant example.
    What speed do you think signals travel over CAT5 ? Now that you looked that up is it faster , slower or the same speed as the microwaves used for satellite communications? The majority of people are not using satellites for their internet access, even on mid hops. The use of optical is a given, I still am not sure why you would not see my example as relevant to the question ?
    If a signal has to reach geosynchronous orbit and back, the round trip takes .28 seconds. Because geosynch orbit is .14 Light second from Earth. That's the trip from straight up to straight back down. That's ignoring the angular route from satellite to satellite back down to Earth back to satellite, etc.

    By comparison a signal sent through fiber optics on the ground around the world (from near side to far side) is a distance of .07 Light seconds.

    But signal speed isn't a factor here. My point is that fiber optics and satellite communication both use light. Whereas cat5 cable uses electrical current. Which is susceptible to random perturbations. I don't know if that would make a difference, but I suspect it would.

    Transatlantic telegraph wires would be a good comparison. Anyone know what sort of requirements they have?
    Reply With Quote  
     

  15. #14  
    Forum Professor leohopkins's Avatar
    Join Date
    Dec 2006
    Location
    Dulwich, London, England
    Posts
    1,418
    Quote Originally Posted by Numsgil
    Quote Originally Posted by (In)Sanity
    Quote Originally Posted by Numsgil
    Most of the internet travels along either fiber optics or satellites (satellites are why there's several ms of latency in online games. Speed of light lag getting from ground to geosynch orbit and back). So that's not a relevant example.
    What speed do you think signals travel over CAT5 ? Now that you looked that up is it faster , slower or the same speed as the microwaves used for satellite communications? The majority of people are not using satellites for their internet access, even on mid hops. The use of optical is a given, I still am not sure why you would not see my example as relevant to the question ?
    If a signal has to reach geosynchronous orbit and back, the round trip takes .28 seconds. Because geosynch orbit is .14 Light second from Earth. That's the trip from straight up to straight back down. That's ignoring the angular route from satellite to satellite back down to Earth back to satellite, etc.

    By comparison a signal sent through fiber optics on the ground around the world (from near side to far side) is a distance of .07 Light seconds.

    But signal speed isn't a factor here. My point is that fiber optics and satellite communication both use light. Whereas cat5 cable uses electrical current. Which is susceptible to random perturbations. I don't know if that would make a difference, but I suspect it would.

    Transatlantic telegraph wires would be a good comparison. Anyone know what sort of requirements they have?
    Yes, normal UTP Cat5, that is (Unshielded Twisted Pair cable) can be affected by radio interferance. Co-axial cable is shielded and is less affected by RF interfereance, as is the more expensive Shielded Twisted Pair cable. However fibre optics are not affected by RF but are a lot more expensive. So the question is if you were called to install a network in an office building, would you ivest on huge sums of money to install fibre optic cabling or shielded twisted pair or co-axial?......

    .....Well, coaxial doesnt give much in the way of bandwidth so unless you were going to install a small token-ring network then i wouldnt recommend co-ax. Both shielded twisted pair and fibre-optics provide good bandwidth and good RF shielding, however the cost of buying the cables would far outweigh the benefits of installing the cable in a single office block. So UTP would be the natural choice. If, however you were going to span the cables over a much longer distance, sya a MAN or a WAN, then fibre optics would be the natural choice simply because you'd get longer runs out of them before you'd need a repeater. (Around 2,000km) Add that to the benefit of reduced RF and you can see that the benefits here outweigh the cost because you do not have the cost of installing a repeater every 98 metres along the network. And this is why most telecommunication providers are now replacing their old copper twisted-pair cables with fibre-optics.
    The hand of time rested on the half-hour mark, and all along that old front line of the English there came a whistling and a crying. The men of the first wave climbed up the parapets, in tumult, darkness, and the presence of death, and having done with all pleasant things, advanced across No Man's Land to begin the Battle of the Somme. - Poet John Masefield.

    www.leohopkins.com
    Reply With Quote  
     

  16. #15  
    Suspended
    Join Date
    Apr 2008
    Posts
    2,176
    Quote Originally Posted by leohopkins
    Quote Originally Posted by Numsgil
    Quote Originally Posted by (In)Sanity
    Quote Originally Posted by Numsgil
    Most of the internet travels along either fiber optics or satellites (satellites are why there's several ms of latency in online games. Speed of light lag getting from ground to geosynch orbit and back). So that's not a relevant example.
    What speed do you think signals travel over CAT5 ? Now that you looked that up is it faster , slower or the same speed as the microwaves used for satellite communications? The majority of people are not using satellites for their internet access, even on mid hops. The use of optical is a given, I still am not sure why you would not see my example as relevant to the question ?
    If a signal has to reach geosynchronous orbit and back, the round trip takes .28 seconds. Because geosynch orbit is .14 Light second from Earth. That's the trip from straight up to straight back down. That's ignoring the angular route from satellite to satellite back down to Earth back to satellite, etc.

    By comparison a signal sent through fiber optics on the ground around the world (from near side to far side) is a distance of .07 Light seconds.

    But signal speed isn't a factor here. My point is that fiber optics and satellite communication both use light. Whereas cat5 cable uses electrical current. Which is susceptible to random perturbations. I don't know if that would make a difference, but I suspect it would.

    Transatlantic telegraph wires would be a good comparison. Anyone know what sort of requirements they have?
    Yes, normal UTP Cat5, that is (Unshielded Twisted Pair cable) can be affected by radio interferance. Co-axial cable is shielded and is less affected by RF interfereance, as is the more expensive Shielded Twisted Pair cable. However fibre optics are not affected by RF but are a lot more expensive. So the question is if you were called to install a network in an office building, would you ivest on huge sums of money to install fibre optic cabling or shielded twisted pair or co-axial?......

    .....Well, coaxial doesnt give much in the way of bandwidth so unless you were going to install a small token-ring network then i wouldnt recommend co-ax. Both shielded twisted pair and fibre-optics provide good bandwidth and good RF shielding, however the cost of buying the cables would far outweigh the benefits of installing the cable in a single office block. So UTP would be the natural choice. If, however you were going to span the cables over a much longer distance, sya a MAN or a WAN, then fibre optics would be the natural choice simply because you'd get longer runs out of them before you'd need a repeater. (Around 2,000km) Add that to the benefit of reduced RF and you can see that the benefits here outweigh the cost because you do not have the cost of installing a repeater every 98 metres along the network. And this is why most telecommunication providers are now replacing their old copper twisted-pair cables with fibre-optics.

    Single wire, shielded coax cable can stream simultaneously over 100 channels of video. That is a fair amount of communication.

    The computer industry never pressed it much, because it can be prone to errors, if you use it to that extent, for exacting computer purposes. However you could use it to communicate a lot of data. It was just the low quality of the equipment that was being used, with the single coax cable.

    Heck the cable vision cable to your house was carrying, 100 plus television channels and the Internet.

    Sincerely,


    William McCormick
    Reply With Quote  
     

  17. #16  
    Forum Professor leohopkins's Avatar
    Join Date
    Dec 2006
    Location
    Dulwich, London, England
    Posts
    1,418
    The cable you speak of William, to my knowledge is fibre optic. Indeed co-axial was used in days gone buy for LAN ethernet, using the old BNC connectors, the problem as you suggested was reliability. Bandwidth was also an issue, as was network congestion.

    As I said co-ax cable (i.e 10Base2 or 10Base5) only carried 1Mbps or 10Mbps respectively, which by todays standards isnt an awful lot. Added to that the fact that it only supported baseband signalling and you have something which is pretty slow for the networks we use today. In truth fibre-optics rule but the cost is phenomenal.
    The hand of time rested on the half-hour mark, and all along that old front line of the English there came a whistling and a crying. The men of the first wave climbed up the parapets, in tumult, darkness, and the presence of death, and having done with all pleasant things, advanced across No Man's Land to begin the Battle of the Somme. - Poet John Masefield.

    www.leohopkins.com
    Reply With Quote  
     

  18. #17  
    Suspended
    Join Date
    Apr 2008
    Posts
    2,176
    Quote Originally Posted by leohopkins
    The cable you speak of William, to my knowledge is fibre optic. Indeed co-axial was used in days gone buy for LAN ethernet, using the old BNC connectors, the problem as you suggested was reliability. Bandwidth was also an issue, as was network congestion.

    As I said co-ax cable (i.e 10Base2 or 10Base5) only carried 1Mbps or 10Mbps respectively, which by todays standards isnt an awful lot. Added to that the fact that it only supported baseband signalling and you have something which is pretty slow for the networks we use today. In truth fibre-optics rule but the cost is phenomenal.
    It carried that amount of data, set by some standard, or convenient technology that could be implemented cheaply.

    But a single 75 ohm coax cable from Cable Vision was streaming over 100 simultaneous video channels to a home, over one piece of copper. This is very common knowledge still today. In a few more years you will not be able to prove this, as everything goes digital.

    Now if they used quality equipment, you could get a lot more transmitted over a single cable. But quality is a thing of the past.

    I am not arguing this. I am just highlighting there are two ways to look at it. What was, in some cases and what was in others.

    I know years ago some were outraged when I told them that I was getting streamed over 100 video channels through a single coax simultaneously.

    Because they never knew.



    Sincerely,


    William McCormick
    Reply With Quote  
     

  19. #18  
    Forum Freshman
    Join Date
    Mar 2009
    Posts
    6
    simple answer is yes but you gotta watch out because the cables itself have limitation.

    most of us are using cat5 which is 100m (about 300+ ft).
    Reply With Quote  
     

  20. #19  
    Forum Professor leohopkins's Avatar
    Join Date
    Dec 2006
    Location
    Dulwich, London, England
    Posts
    1,418
    Quote Originally Posted by kelcya
    simple answer is yes but you gotta watch out because the cables itself have limitation.

    most of us are using cat5 which is 100m (about 300+ ft).
    To be precise its 98m.
    The hand of time rested on the half-hour mark, and all along that old front line of the English there came a whistling and a crying. The men of the first wave climbed up the parapets, in tumult, darkness, and the presence of death, and having done with all pleasant things, advanced across No Man's Land to begin the Battle of the Somme. - Poet John Masefield.

    www.leohopkins.com
    Reply With Quote  
     

Bookmarks
Bookmarks
Posting Permissions
  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •