Will@cup.portal.com (Will E Estes) (02/26/90)
I want to connect 5 people in an Ethernet LAN, and the number of connections could grow to a total of 20 by the end of the year. I want to use Ethernet because of the speed, and I want something that is both inexpensive and flexible enough to allow us to grow. My questions are: 1) If I go the 10BaseT route, what are the pieces of hardware I need to buy, and what are the costs? I realize I need 5 10BaseT cards and the RJ45. What else? 2) I have heard some people express concerns that the 10BaseT vendors manufactured their cards before the spec was finished. Is this a concern, or did the spec that was used to build cards end up being the final spec? In particular, will I be safe buying 3COM's 10BaseT cards? 3) Are there any potential compatibility problems in putting together a heterogeneous network of PCs, Macs, Suns, and DEC VMS workstations? Thanks, Will (sun!portal!cup.portal.com!Will)
haas@cs.utah.edu (Walt Haas) (02/26/90)
In article <27314@cup.portal.com> Will@cup.portal.com (Will E Estes) writes: >I want to connect 5 people in an Ethernet LAN, and the number of connections >could grow to a total of 20 by the end of the year. I want to use >Ethernet because of the speed, and I want something that is both inexpensive >and flexible enough to allow us to grow. My questions are: > >1) If I go the 10BaseT route, what are the pieces of hardware I need to >buy, and what are the costs? I realize I need 5 10BaseT cards and the >RJ45. What else? You need an active hub to plug all machines into. The hub needs a port with a repeater for each machine. You will hence need to expand the hub as you expand your network. >2) I have heard some people express concerns that the 10BaseT vendors >manufactured their cards before the spec was finished. Is this a >concern, or did the spec that was used to build cards end up being >the final spec? In particular, will I be safe buying 3COM's 10BaseT >cards? It is definitely a concern at this time, Maybe not in a year or so. In particular, the spec isn't even finished - we're all just betting tht this draft won't be sent back for revision before final approval. >3) Are there any potential compatibility problems in putting together >a heterogeneous network of PCs, Macs, Suns, and DEC VMS workstations? Software compatibility of course - what protocol(s) are you going to run? Cheers -- Walt
John_Robert_Breeden@cup.portal.com (02/27/90)
>1) If I go the 10BaseT route, what are the pieces of hardware I need to >buy, and what are the costs? I realize I need 5 10BaseT cards and the >RJ45. What else? > Hubs (repeaters - NOT - concentrators). 10baseT is star configuration ONLY and the total of the hardware IS more expensive than thinnet. ie: cost of 10baseT cards+hubs+wire/#ports > 10base2 cards+wire/#ports. >2) I have heard some people express concerns that the 10BaseT vendors >manufactured their cards before the spec was finished. Is this a >concern, or did the spec that was used to build cards end up being >the final spec? In particular, will I be safe buying 3COM's 10BaseT >cards? > Those vendors that held out support for 10baseT (read Synoptics and Cabletron) liked to harp on that . The truth is that the draft was well enough defined 2 years ago to assure that product produced then would at least be complient (ie: would work with) final standard product. Examples of this are AT&T, HP and UB's products - all designed and built 2 years ago and all interwork with product today. For the most part it was a smoke screen to allow vendors to continue to sell propriatary product. It's a moot point now - everybody and their brother (and the family dog) has announced or released 10baseT product. By the way, as of today, 10baseT is STILL a draft, not a standard. It's kind of funny that the vendors who six months ago screamed loudest about "compatability" have announced product before 10baseT is a "standard". Pity the poor folks that now have to "migrate" (ie:throw away) their propriatory product to conform to the 10baseT draft. >3) Are there any potential compatibility problems in putting together >a heterogeneous network of PCs, Macs, Suns, and DEC VMS workstations? If you can do it with 802.3 - you can do it with 10baseT. ----- Though 10baseT's installation cost is greater up front than 10base2 (the hubs 'il get ya every time) - it does have it's benefits. 1. Can use pre-existing TP building wire (save the cost of coax). 2. Isolation of bad nodes (star configuration instead of bus). Same wiring configuration as phones - using same wiring paths and wire closets as phones, makes it easier to handle. 3. Intelligence. Some hubs let you turn on and off ports from a remote console and set error levels etc. Some "super-intellegent" hubs even do this automatically - without human intervention (ie; AT&T). inside - like that Bill? (-: 4. Easier to find wire faults (ever try to find a break in coax?). 5. If you miss the boat on 10baseT, not to worry, 10baseF is "acomin' down the slip" (-: john_robert_breeden@cup.portal.com "Please EMail Flames directly, I send them to Mom"
dan@scooter.rosemount.com (Dan Messinger) (03/02/90)
>1) If I go the 10BaseT route, what are the pieces of hardware I need to >buy, and what are the costs? I realize I need 5 10BaseT cards and the >RJ45. What else? First, you need the chassis that the 10BaseT cards go into. 3Com has a box that the call a MultiConnect repeater. I have one here. It holds about 15 cards, I think (you don't really expect me to walk to the other end of the building to verify that, do you?) 3Com also has a device that they call a PairTamer. It converts thinnet to 10BaseT. If you use a thinnet card in the MultiConnect box, then you need a pair of PairTamers for each TP wire. I hear that 3Com will soon have a card for the MultiConnect that is 10BaseT, so then you would only need a single 10BaseT converter on the far end of the TP wire. Note that at the far end of the TP, the coverter gives you thinnet. So if you have workstations that are reasonably close to each other, you can hang several on the same thinnet line. This will reduce the number of 10BaseT lines that you will need. >2) I have heard some people express concerns that the 10BaseT vendors >manufactured their cards before the spec was finished. Is this a >concern, or did the spec that was used to build cards end up being >the final spec? In particular, will I be safe buying 3COM's 10BaseT >cards? My understanding is that the approved standard (which is basically decided upon) will match what 3Com has been selling as 10BaseT. (3Com made a gamble, and it looks like they won) dan@scooter.rosemount.com
John_Robert_Breeden@cup.portal.com (03/05/90)
>>1) If I go the 10BaseT route, what are the pieces of hardware I need to >>buy, and what are the costs? I realize I need 5 10BaseT cards and the >>RJ45. What else? > >First, you need the chassis that the 10BaseT cards go into. 3Com has a >box that the call a MultiConnect repeater. I have one here. It holds >about 15 cards, I think (you don't really expect me to walk to the other >end of the building to verify that, do you?) 3Com also has a device that >they call a PairTamer. It converts thinnet to 10BaseT. If you use a thinnet >card in the MultiConnect box, then you need a pair of PairTamers for each >TP wire. I hear that 3Com will soon have a card for the MultiConnect that DANGER WILL ROBENSON, DANGER!!! The 3Com pairtamer IS NOT A 10BASET DEVICE!!! it is a BALUN - A PASSIVE DEVICE, if you bother to read the 10baseT draft, you'll find the BALUNS ARE A NONO!!! <--tech. term. To hook a thick/thinnet to TP you need an AUI (UB calles it a TPAU), it's an ACTIVE device with all kinds of goodies like receiver, transmitter, signal processor etc. It adds the nessessary precomp and jitter to the signal among other things and it's what gives you the 100 meters in 4pair or 25 pair. Is the Pairtamer cheaper than an AUI - you bet!, does it work - it might, is it 10baseT? - NO WAY JOSE!! >is 10BaseT, so then you would only need a single 10BaseT converter on the far >end of the TP wire. > >Note that at the far end of the TP, the coverter gives you thinnet. So if >you have workstations that are reasonably close to each other, you can >hang several on the same thinnet line. This will reduce the number of 10BaseT >lines that you will need. NO, NO, NO, NO, NO! Don't call nodes hooked up in a daisy chain 10baseT, it ain't!!! READ THE DRAFT! IT MUST BE IN A STAR CONFIGURATIONS. NODES THAT HAVE THINNET GO TO AN AUI ADAPTER (active device NOT A BALUN) and then DIRECTLY TO A HUB (I hate to tell you this, a 10baseT hub must be A ACTIVE REPEATER - NOT A CONCENTRATOR!). >>2) I have heard some people express concerns that the 10BaseT vendors >>manufactured their cards before the spec was finished. Is this a >>concern, or did the spec that was used to build cards end up being >>the final spec? In particular, will I be safe buying 3COM's 10BaseT >>cards? > >My understanding is that the approved standard (which is basically decided >upon) will match what 3Com has been selling as 10BaseT. (3Com made a gamble, >and it looks like they won) If the idea of using baluns and daisy-chaining nodes with TP is is matching "what 3Com has been selling as 10BaseT" - that's just, well WRONG. In defense of 3Com they have announced intended support of 10baseT and have, in fact, released a 10baseT version of the EtherLink II. They where one of the participents in the 10baseT "let's hook it alltogther" 10baseT interoperability demonstration at Interop in Boston last month along with AT&T, UB, Nextworth, Cabletron, HP and others (notice lack of one of the big TP vendors). When I talked to a spokesman at 3Com about six months ago, 3Com stated they "would release 10baseT product when the draft was a standard" - along with the rest of the world. So the bottom line is if you use the solution above, you are NOT 10baseT but instead are running a PROPRIATARY 3COM SOLUTION - nothing wrong with that as long as you are prepared to live with the fact that you may (read most likely) will not be able to use other vendor's true 10baseT product in the same network. BTW, this is NOT a flame of 3Com, they have been an active, and I under- stand, supportive member of the 10baseT committee. They where in fact, one of the first companies to offer 10meg support on TP long before the WAS a 10baseT committee (that's where the Pairtamer/balun came from). Compair that to another Silicon Valley TP vendor who will remain un-named (it starts with an "S") who has used their seat on the committee to confuse and delay the release of the draft to protect their propriatery position. john_robert_breeden@cup.portal.com
mbt@bridge2.ESD.3Com.COM (Brad Turner) (03/06/90)
John_Robert_Breeden@cup.portal.com writes: >>>1) If I go the 10BaseT route, what are the pieces of hardware I need to >>>buy, and what are the costs? I realize I need 5 10BaseT cards and the >>>RJ45. What else? >> >>First, you need the chassis that the 10BaseT cards go into. 3Com has a >>box that the call a MultiConnect repeater. I have one here. It holds >>about 15 cards, I think (you don't really expect me to walk to the other >>end of the building to verify that, do you?) 3Com also has a device that >>they call a PairTamer. It converts thinnet to 10BaseT. If you use a thinnet >>card in the MultiConnect box, then you need a pair of PairTamers for each >>TP wire. I hear that 3Com will soon have a card for the MultiConnect that >DANGER WILL ROBENSON, DANGER!!! The 3Com pairtamer IS NOT A 10BASET DEVICE!!! >it is a BALUN - A PASSIVE DEVICE, if you bother to read the 10baseT draft, [stuff deleted] >If the idea of using baluns and daisy-chaining nodes with TP is is matching >"what 3Com has been selling as 10BaseT" - that's just, well WRONG. [stuff deleted] >So the bottom line is if you use the solution above, you are NOT 10baseT >but instead are running a PROPRIATARY 3COM SOLUTION - nothing wrong with >that as long as you are prepared to live with the fact that you may (read >most likely) will not be able to use other vendor's true 10baseT product >in the same network. The Pairtamer has NEVER been marketed as a 10BaseT product! It is marketed as a 10Base2 product which is exactly what it is. The Pairtamer can and will coexist with other 10BaseT products. For example the configuration below: --- thinnet XXX repeater that supports 10BaseT and 10Base2 ][ 2 pair of TP wiring = 1 pair of TP wiring @ Pairtamer module ## end node (either a 10BaseT enet interface or a 10BaseT transceiver) -------XXXXXXXX----@===================@----XXXXXXXX------- ][ ][ ][ ][ ][ ][ ][ ][ ][ ][ ][ ][ ][ ][ ][ ][ ][ ][ ][ ][ ][ ][ ][ ][ ][ ][ ][ ][ ][ ][ ## ## ## ## ## ## Where the confusion is going to happen is when folks do not understand the difference between the Pairtamer and 10BaseT devices. You CANNOT plug an RJ-45 into a Pairtamer on one end and a 10BaseT device on the other end (well actually you can, the plugs will fit together, but it is not going to function.) In some instances the Pairtamer does make more sense than 10BaseT (e.g. cost per port, # of pairs required, repeater penalty.) Before this turns into a commercial; call 800-NET-3COM if you want more information. -brad- v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v Brad Turner |2081 Shoreline Blvd.|(415) 969-2099 ext 217 | I speak for myself 3Com Corp. |Mtn. View, CA 94043 |mbt@bridge2.ESD.3Com.Com| NOT for my employer -- v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v^v Brad Turner |2081 Shoreline Blvd.|(415) 969-2099 ext 217 | I speak for myself 3Com Corp. |Mtn. View, CA 94043 |mbt@bridge2.ESD.3Com.Com| NOT for my employer
root@dosadi.UUCP (Dave Edick) (03/06/90)
dan@scooter.rosemount.com (Dan Messinger) writes: > > My understanding is that the approved standard (which is basically decided > upon) will match what 3Com has been selling as 10BaseT. (3Com made a gamble, > and it looks like they won) > The 3Com PairTamers are NOT 10BaseT compatible. They're really nothing other than a balun (a transformer). 3Com will supposedly be introducing true 10BaseT compatible hardware soon. Basically, neither 3Com nor Synoptics "won" on the 10BaseT spec. The spec is much closer to Synoptics, but the commitee made changes which make Synoptics hardware incompatible with the "final" spec. From what I've heard recently, the "final" spec won't really be approved until July at the earliest. Anyone know the actual approval date? By the way, there's a product called EtherNext (forgot the company name), which the manufacturer guarantees will be compatible with the final spec (they'll ship an upgrade if the spec changes). Don't know anything about them other than what I read in the literature I got a few days ago. ------- /Dave Edick/ {pacbell,rencon,dumbcat,sorinc}!esfenn!dosadi!root "I can't go to a single solar system without apologizing for you." Star Trek, The Next Generation - Deja Q
pat@hprnd.HP.COM (Pat Thaler) (03/20/90)
> > You should NOT be concerned about the standard changing...here is the scoop > > The 10BASE-T Committee began its work in Aug 87, by reviewing the eight > formal proposals submitted as part of the preliminary proceedings. > There were distilled into a single draft standard and submitted to the > IEEE for review and comment. 2 1/2 years (and several drafts) later, the > standard has passed confirmation vote and is now waiting approval by the > Technical Committee on Computer Communications and then it will be ready for > final approval by the IEEE Standards Board. At this stage, no technical > changes to the spec are expected and several IEEE committee members have > introduced and are shipping their 10BASE-T products (e.g. 3Com). Final > publication will occur in September 1990. > > -Marianne Cohn > ---------- The 10BASE-T draft has passed voting by IEEE 802.3. While I would agree with Marianne that at this point I don't _expect_ technical changes, the draft is not yet a standard and it is _possible_ that technical changes could arise from the TCCC balloting. TCCC ballotting should be over by mid-May and 10BASE-T will meet to resolve ballot comments the week of May 21. If all goes well, IEEE 802.3 will review our comment resolution at the July meeting and 10BASE-T will be forwarded to the Sept 28 meeting of the IEEE Standards Board, the final step to becoming an IEEE standard. 10BASE-T is also being submitted to JTC-1 for balloting as a proposal for an international standard. Pat Thaler Chair of IEEE 802.3