--------------- FIDO MESSAGE AREA==> TOPIC: 193 LAN Ref: DEG00000 Date: 10/10/96 From: MUFUTAU TOWOBOLA Time: 07:59pm \/To: SHANE CANARD (Read 9 times) Subj: NE1000 |---------------|Shane Canard wrote: SC>Could somebody let me know the advantages of NE2000 over the NE1000 are SC>It will only be a tiny (4 Node) Network, Peer-Peer, for file sharing and SC>printing using Win 3.11 and Win '95 on Thin Coax (w/Bnc). You will get more IRQ and address selection on the NE2000 because they are always 16bit cards, and it will be faster than the NE1000 which is just an 8bit card and are only good during the life of XT computers... unless you already have the NE1000, but if you are gonna go buy them, then you should get the NE2000 for speed and all the goodies expolained above... --- WILDMAIL!/WC v4.12 * Origin: Systematic BBS, Bronx, NY (718) 716-6198 (1:278/111.0) --------------- FIDO MESSAGE AREA==> TOPIC: 193 LAN Ref: DEG00001 Date: 10/09/96 From: JEFF DUNLOP Time: 08:28pm \/To: DAVID MOHORN (Read 9 times) Subj: Subnetting IP Addresses DM> Does anyone have any good explanations of how to subnet and how to DM> figure out what nodes are on a particular subnet? The following is the body of a technote I created some time ago: Subnetting is mandatory if you have multiple segments connected to an internetwork. The multiple segments are connected to a router, and the router must be able to determine by inspecting the address which wire the packet goes to. Netware does routing, by the way. This entire discussion assumes a Class C address (xxx.yyy.zzz.nnn). Subnetting is the arbitrary division of bits in the 32-bit address. You have 32 bits, and any number of them can be the network ID, with the balance being the host (node) ID. Since the first 24 bits are handed to you by the ISP, you have only 8 bits to work with. You can split the 8 bits any way you want between the subnet and host fields, for instance 3 bits subnet and 5 bits host. In the 3/5 example, there are 8 possible subnets, and 32 possible hosts for each subnet. Naming rules -- no field is 0 or all 1's -- reduce the actual usable number of subnets to 6 and hosts to 30. Think of the subnet as an additional dotted field for the moment. The first node of the first subnet would thus be xxx.yyy.zzz.1.1. To arrive at the actual 4D dotted notation, then, you left-shift the subnet ID 5 places and add the node. 1 << 5 is 32, and 32 + 1 is 33, so the first address on segment one would be xxx.yyy.zzz.33. The last possible address would be .62. The first address on the second segment would be .65, and the last .94. The subnet mask must be entered into every device and program that delivers IP packets, including the Netware server, the router, and every client application's networking protocol. The mask is nothing but a series of binary 1's and 0's. If a particular bit is part of the network (and subnet) ID, it's a 1. If it's part of the host address, it's a zero. With a 3-bit subnet in the last field, the mask is thus 255.255.255.224. 255 is obviously binary 11111111. 224 is binary 11100000. The table for subnet/host counts is: Mask bits Mask Mask Available Hosts per Total Hosts (Binary) (Decimal) Subnets Subnet Available 1 10000000 128 0 N/A 0 2 11000000 192 2 62 124 3 11100000 224 6 30 180 4 11110000 240 14 14 196 5 11111000 248 30 6 180 6 11111100 252 62 2 124 7 11111110 254 126 0 0 8 11111111 255 254 0 0 The number of bits you use is completely dependent on the maximum number of nodes you have or plan to have on the wire, tempered by the maximum number of wires you plan to run. The nasty part of this is that if you change your mind on one of the maximums -- nodes/wire or wires -- you get to visit every client and router to update its address and mask, and you get to keep straight exactly which subnet every node is now on. This is the single most time-consuming maintenance aspect of statically assigned TCP/IP. Jeff --- GoldED/2 2.42.G0615 * Origin: DB/Soft Online - Sacramento, CA (916)927-2349 (1:203/16) --------------- FIDO MESSAGE AREA==> TOPIC: 193 LAN Ref: DEG00002 Date: 10/10/96 From: BRUCE LANE Time: 10:45pm \/To: SCOTT PARKS (Read 9 times) Subj: Re: CSU standards? Scott Parks woke half the neighborhood at 03:12 by yelling at Bruce Lane about CSU standards?... SP> Timing is either 56k or 64k ... is what I assume you mean. Correct. Again, in most 'switched' environments or 'DDS' services, the local telco provides timing. SP> Don't remember off hand .... RE-PC have any ? I remember a couple of 3510's hanging around. They've been gathering dust for a while, so you might actually get them at a reasonable price. ;-) ==Bruce ... Press to test... Release to detonate! --- Blue Wave/RA v2.21 * Origin: No drones allowed... (1:343/272) --------------- FIDO MESSAGE AREA==> TOPIC: 193 LAN Ref: DEG00003 Date: 10/11/96 From: CHRIS MADDOCK Time: 08:24pm \/To: KURT HILL (Read 8 times) Subj: coax headaches On 07 Oct at 17:37, Kurt Hill of 1:202/107 wrote to Eric Smith: KH> True COAX? Skin effect? I have heard (from reputable persons) KH> that braided is better. If, by skin-effect, you mean that the signal KH> travels on the surface of the conductor, then I would ask you to explain KH> this in some more detail, please. [.................] KH> I invite *ALL* FidoNet denizens to contribute and decide, once and KH> for all, which is better in the braided Vs. solid war. Bring me your KH> specs, your kinks and your impedences! "Skin Effect" with regard to high frequency electrical signals is (in a nutshell), caused by the flow of electrons along a wire causing a magnetic field. One of the interesting effects of an inductor is that during any current change along the wire, the changing magnetic field produced by the current, is such an orientation that the changing field will produce a voltage in the conductor that opposed the direction of original voltage change. (Substitute current for voltage in the preceeding sentence and it hold true as well.) A straight wire has inductance. Strange ? It's true ! Think about it. Pass a current down a straight piece of wire. It sets up a magnetic field which produces a voltrage in the wire that opposes the change. Ergo - it is an inductor. A condition that absolutely *must* hold true if any antenna is to work. Now, the magnetic field is greatest in the centre of the conductor. It's the area of greatest effect of the magnetic field. Because of this, the magnetic field has the greatest effect at the *centre* of the conductor with the effect diminishing towards the outside of the same conductor. As a result, at very high frequencies there is a pronounced "skin effect" where the current is flowing towards the outside of the conductor as the inductance effect is lowest there. At Super High Frequencies (SHF), the skin effect is so much that "waveguides" are used. A wave guide at say microwave frequencies (From memory 10-30GHz) is constructed in such a way that the skin effect is "guided" up a tuned cavity with a "horn" forming the interface from the "conductor" (waveguide) and free air (at the point of transmission) or a "collector" (a tuned - normally 1/2 wave dipole antenna). Skin effect at 10MHz is negligible. At 100MHz it is significant and at a 1000MHz it is *very* significant. Depending on the application of course. Taking all this into account, it is better to have many fine conductors than a single conductor, at very high frequencies. It results in a very much greater surface area for the current to flow and therefore less resistance in the conductor. Litz wire is one prime example for instance. The effect, as noted above, is negligible at 10MHz. i.e. Foil covered coax is fine. It's just so godamned difficult to terminate - and *stay* terminated (with a connector). The difference between single conductor and multi-strand is small. The ability to easily fracture a single conductor however is critical. Multi-strand is totally superior in an aplication that requires flexability - such as network coax cable (Thin-net). Regards, Chris Maddock chrism@softtech.brisnet.org.au --- Msged/386 4.00 * Origin: Diagnostic CBBS - BN QL AU - (3:640/302) --------------- FIDO MESSAGE AREA==> TOPIC: 193 LAN Ref: DEG00004 Date: 10/10/96 From: ERIC SMITH Time: 02:57pm \/To: KURT HILL (Read 8 times) Subj: coax headaches KH> True COAX? Skin effect? I have heard (from reputable persons) that >braided is better. If, by skin-effect, you mean that the signal travels n >the >surface of the conductor, then I would ask you to explain this in some ore >detail, please. My understanding of it is that voltages (potential >differences) cause a current flow -- the actual flow is measured in feet t >inches per second (quite slow), but the *amount* of the flow, which is >regulated by the voltage, can change nearly instantly. The only "skin" >effect >I am familiar with is related to static charges, which are essentially >surface To best look at the skin effect we will use CATV as an example (Since most everybody has it and can test this for themselves. Please excuse my crude description, a CATV engineer can explain it much better. Ever wonder how the cable company can send so many channels over the same cable? And why they have a advantage over the phone companies in bringing services to the homes? It's because of the COAX cable that they use. The little Copper wire that sticks out of the CATV cable s all that is really needed. If you have one with a long enough extension on it, unplug it from your TV and slowly just push it back in without screwing or touching the outside parts. You will see you TV signal come to life. It works this way because of the Skin effect. The electrical signal is sent through it's core. Each channel has it's own frequency, which moves out slightly from the core towards the surface (Note I could have the core and surface backwards, but I believe the charge is in the core). The channel you are watching goes back on that frequency which registers on the equipment at the Cable company. That is how they are able to tell what channel you are watching, and how channels like some pay per views work (Watch it long enough and you get billed). Now days, the equipment at the cable company can detect and sort multiple signals, which is what allows you to have multiple TV's showing different channels at the same time. In the old days, it didn't work so well. Now the differences in cable is this. 10Base-T cable or CAT-5 rather, is braided. It does this because of electro magnetic interference. A strong radio r electrical signal on one side of it will build up a uneven charge causing heavy static or signal bleed. The wires braid so that when they switch sides, the other wire gets the same charge, and so they cancel that charge out. The more twist the better. CATV cables have the same problem, except there is only one wire in it, so it must be shielded. The cheep stuff you get in stores has a aluminium foil type of material inside the black rubber, then a white hard rubber, then the single wire. The better cables have a raid of wires in place of the aluminum (Which is probably what is being referred to) and often has a type of jelly in with it so that the cable can bend and move without breaking those brades. I hope it answers your question for you. Try the thing with the CATV at home, it's kind of fun. --- * QMPro 1.53 * Logic Error in CLINTON.SYS - Truth table missing. --- WILDMAIL!/WC v4.12 * Origin: Aspencade BBS (1:226/0.0) --------------- FIDO MESSAGE AREA==> TOPIC: 193 LAN Ref: DEG00005 Date: 10/03/96 From: ERIC MCCORMICK Time: 02:01pm \/To: LUIS MANTEROLA (Read 8 times) Subj: Numbering EM> bnc connectors on one of the cards is defective. EM> It does not matter much EM> to me. It is working, that is that. LM> So, there could be some other things such as the cards LM> were configured for UTP and not for coax. Or the coax LM> was not grounded on one of it's terminations. I seriously doubt it. I do not have the equipment to put something in line and test it, however, I repetivly checked the CMOS settings on the cards. In fact, I had to turn them back to UTP mode to get that to work when I cross wired the cable. That is what leads me to beleve it is one of the cards. Probably something as simple as a soldered contact, or a fried resistor. --- Maximus/2 3.00 * Origin: The NIMBY Zone (1:207/406) --------------- FIDO MESSAGE AREA==> TOPIC: 193 LAN Ref: DEG00006 Date: 10/12/96 From: SHANE CANARD Time: 03:46pm \/To: ALL (Read 8 times) Subj: Packet Driver for Arcnet NICS Greetings, Does anybody have a packet driver for SMC Arcnet Network Cards suitable for NEOS 1.0 ? I think that any packet driver for these cards should work, shouldn't they ? If anybody does, could you please UUENCODE it and email or netmail it to me at either 3:640/302.5 or shane@softtech.brisnet.org.au Thanx, Shane "Duck" Canard [ shane@softtech.brisnet.org.au ] -== Ya Gotta Be Quackers to Survive ==- PGP fingerprint = 7A 1F B8 BD 1C 8D 5D E9 FF 56 44 A3 1B 7F 51 75 ... Data convinces the Coke machine that Pepsi is better. --- PPoint 2.00 * Origin: The Ducks Cyberpond, Tamborine, Aust (3:640/302.5) --------------- FIDO MESSAGE AREA==> TOPIC: 193 LAN Ref: DEGQ2002 Date: 10/12/96 From: THOMAS KAUL Time: 09:33pm \/To: ALL (Read 9 times) Subj: CONNECTING Anybody know why one computer sits at the "searching cards" for like 5 seconds, then it won't connect to tho others? Adam Kaul --------------- FIDO MESSAGE AREA==> TOPIC: 193 LAN Ref: DEH00000 Date: 10/13/96 From: BEN BERGERON Time: 02:33am \/To: ALL (Read 9 times) Subj: NW 3.12 & ICLASS Chello!. Tad problem. One of the admins added me to be a PSERVER op the other day, now whenever anyone prints to ANY QUE, in PCONSOLE, it either says SYSOP or BERGERONB (Me :).... it's really wacked out because if one checks the status of the QUEUE, ie, hit on the Job, it shows who is really sending the job in the client? field. I suggested he run Binfix, but he didn't want to because he said he would have to down the network. Typical, if it still works, no one cares to fix it befor it does break. Also, we are running into a tad little problem (besides the above). If any of the 30 some P75's running win 3.1 off of local drives, sit in their screen saver for a period of time, I think around 10 - 20 min, they lock up. I was reading about somthing simular to this on Novells web page. The black screen of death. I dl'ed what I think are the new drivers that are to fix this problem and a few others but I don't want to change anything untill I am sure they are the right ones and this is really the problem. And lastly, (For Now :) .... on the same systems with the screen of death problems, we have Netscape 2.02? installed. Locally of course :). The problem is When there are more then say 10 - 15 people using Netscape, IT GOES SO SLOW!, kinda makes 300 Baud look like High Speed!. The pipe so to speak is equivilent to ISDN with Both B channels. I understand that with upto 30 people going at once threw a small little pipe, things will go slow, but people are not always using the net, they DO STOP and read :). Is there any tips to tweaking such a setup? ... I beleave most of the slowdown is coming from downtown, but if there is anything we can do to help, I would like to know. Closing, Does windows REALLY NEED a swap file?? :) --- 6b416e55724565446549