From michael at priori.net Tue Sep 9 14:54:15 1997 From: michael at priori.net (Michael Dillon) Date: Tue, 9 Sep 1997 11:54:15 -0700 Subject: too many routes In-Reply-To: <199709091815.NAA18919@charon.milepost.com> References: from "Michael Dillon" at Sep 9, 97 10:05:27 am Message-ID: >We are at a point that we can pretty much nearly fill a /19 and will be >past that point by the end of the year. Beyond that, I would think that >when we need to keep coming back for address space, we would do it in >increments of /19. The problem is that although you know your goals and you also know that you WILL reach your goals, there are other companies out there with wildly inflated goals that will never even come close to reaching them. The registry folks have to cut through the bullshit and try to avoid delegating space to companies who don't really need the space and will never use it. >I'm wondering if the policies are not being counter productive due to the >apparently opposing nature of IP space vs route space (e.g. the one where >people have to take more IP space to be routed, because of route filters, >because of too many routes, because IP space is so fragmented, because IP >space needs to be conserved, because everyone needs to get more IP space, >becayse of route filtering, because of so many routes, because ...). You are quite right, it's a tricky balancing act. And many of us do believe that the right balance has not been struck yet and that allocation policies need to be modified a bit to accomodate the current realities. >I was given some other mailing lists for dealing with those issues by someone. >I'll be subscribing there. Hopefully the NAIPR list was one of them because that is where the new policies will likely be hashed out. Send a subscribe message to naipr-request at arin.net and before posting anything, review the background material at http://www.arin.net including the list archives. At ISPCON 3 weeks ago, Kim Hubbard announced that 50 organizations had already applied for ARIN membership. >Or is this listed on a web page? If not, I'll offer to host just such a >web page if I can get a pretty much complete list (I'll even write HTML to >get it going). But I'm sure most everyone here has some access to some >web space somewhere. I think that the ISP info pages hosted at http://www.nanog.org already has a pretty good collection of pointers to mailing lists. ******************************************************** Michael Dillon voice: +1-650-482-2840 Senior Systems Architect fax: +1-650-482-2844 PRIORI NETWORKS, INC. http://www.priori.net "The People You Know. The People You Trust." ******************************************************** From jwkckid1 at ix.netcom.com Tue Sep 9 12:11:46 1997 From: jwkckid1 at ix.netcom.com (Jeff Williams) Date: Tue, 09 Sep 1997 17:11:46 +0100 Subject: too many routes References: from "Michael Dillon" at Sep 9, 97 10:05:27 am Message-ID: <34157542.59CC@ix.netcom.com> Michael, Michael Dillon wrote: > > >We are at a point that we can pretty much nearly fill a /19 and will be > >past that point by the end of the year. Beyond that, I would think that > >when we need to keep coming back for address space, we would do it in > >increments of /19. > > The problem is that although you know your goals and you also know that you > WILL reach your goals, there are other companies out there with wildly > inflated goals that will never even come close to reaching them. The > registry folks have to cut through the bullshit and try to avoid delegating > space to companies who don't really need the space and will never use it. Well if there were reasonable paramaters set, determining weather or not the requests for allocations would be reasonable. RFC 2050 is insuficient for this perpose, yet is the backdrop or primary document that NAPIR is using as a guideline. Bad policy. > > >I'm wondering if the policies are not being counter productive due to the > >apparently opposing nature of IP space vs route space (e.g. the one where > >people have to take more IP space to be routed, because of route filters, > >because of too many routes, because IP space is so fragmented, because IP > >space needs to be conserved, because everyone needs to get more IP space, > >becayse of route filtering, because of so many routes, because ...). > > You are quite right, it's a tricky balancing act. And many of us do believe > that the right balance has not been struck yet and that allocation policies > need to be modified a bit to accomodate the current realities. A bit! That is an understatment at best. > > >I was given some other mailing lists for dealing with those issues by someone. > >I'll be subscribing there. > > Hopefully the NAIPR list was one of them because that is where the new > policies will likely be hashed out. Send a subscribe message to > naipr-request at arin.net and before posting anything, review the background > material at http://www.arin.net including the list archives. At ISPCON 3 > weeks ago, Kim Hubbard announced that 50 organizations had already applied > for ARIN membership. Yep. And this announcment is premature until some of the allocation pollicies are better worked out. > > >Or is this listed on a web page? If not, I'll offer to host just such a > >web page if I can get a pretty much complete list (I'll even write HTML to > >get it going). But I'm sure most everyone here has some access to some > >web space somewhere. > > I think that the ISP info pages hosted at http://www.nanog.org already has > a pretty good collection of pointers to mailing lists. > > ******************************************************** > Michael Dillon voice: +1-650-482-2840 > Senior Systems Architect fax: +1-650-482-2844 > PRIORI NETWORKS, INC. http://www.priori.net > > "The People You Know. The People You Trust." > ******************************************************** Regards, -- Jeffrey A. Williams DIR. Internet Network Eng/SR. Java Development Eng. Information Eng. Group. IEG. INC. Phone :913-294-2375 (v-office) E-Mail jwkckid1 at ix.netcom.com From jlewis at inorganic5.fdt.net Tue Sep 9 23:41:50 1997 From: jlewis at inorganic5.fdt.net (Jon Lewis) Date: Tue, 9 Sep 1997 23:41:50 -0400 (EDT) Subject: too many routes In-Reply-To: Message-ID: [moved from nanog to a more appropriate forum] On Tue, 9 Sep 1997, Michael Dillon wrote: > > >I would think the latter. If so, would anyone know why it is that the > >backbone providers are so resistant to giving out blocks to do this? > > If backbone operators do not efficiently allocate IP space to their > downstream customers then the next time they need additional IP space they > will not be able to get any. Everyone in the food chain has to operate > under the same policies of justifying IP space based on need and using the > space efficiently. There is more info on this at http://www.arin.net and in > particular you should have a look through the recommended reading section > especially any documents relating to CIDR. How do the rules of rfc2050 apply though when you have a hierarchy of providers and customers, where often the customer is a provider? i.e. picture the food chain I'm part of. UUNet provides a T1 and 2 /20 blocks to FDT (an ISP). Should UUNet give FDT address space based on the 25%/50% rule or the "slow-start" procedure? FDT provides T1 service to several smaller ISP's. Do we allocate space for them using the 25/50 rule or the slow-start rule? These smaller ISP's that feed from us have customers using multiple IP's (selling web space to others). Do they [the smaller ISPs] assign space to their customers using 25/50 or slow-start? It seems the rules are either too simple for the real world or were written with the idea that there are only providers and end users, and that nobody would be both. ------------------------------------------------------------------ Jon Lewis | Unsolicited commercial e-mail will Network Administrator | be proof-read for $199/message. Florida Digital Turnpike | ______http://inorganic5.fdt.net/~jlewis/pgp for PGP public key____ From jtk at titania.net Wed Sep 10 02:53:05 1997 From: jtk at titania.net (Joseph T. Klein) Date: Wed, 10 Sep 97 02:53:05 Central Daylight Time Subject: too many routes References: Message-ID: A system must be established that will NOT keep new entities from entering the market. I would rather see a world with 30 or 40 backbones, vs. a world with one. The concept of upstream providers providing address space does not work for organizations seeking to establish national or international networks. If I am to build a network to 5 major meet points and contract with UUNET, and MCI for transit, why should I put up with the InetrNIC or ARIN telling me that I must get a block of addresses from an upstream??! How the do you expect to establish new backbones without a loophole for new providers to get a /19 and an ASN? Even to anty up to do this, we are in the 1 Million $$$ range. Come on registry people ... give the open market model a chance. --- On Tue, 9 Sep 1997 23:41:50 -0400 (EDT) Jon Lewis wrote: > > > > >I would think the latter. If so, would anyone know why it is that the > > >backbone providers are so resistant to giving out blocks to do this? > > > > If backbone operators do not efficiently allocate IP space to their > > downstream customers then the next time they need additional IP space they > > will not be able to get any. Everyone in the food chain has to operate > > under the same policies of justifying IP space based on need and using the > > space efficiently. There is more info on this at http://www.arin.net and in > > particular you should have a look through the recommended reading section > > especially any documents relating to CIDR. > > How do the rules of rfc2050 apply though when you have a hierarchy of > providers and customers, where often the customer is a provider? i.e. > picture the food chain I'm part of. UUNet provides a T1 and 2 /20 blocks > to FDT (an ISP). Should UUNet give FDT address space based on the 25%/50% > rule or the "slow-start" procedure? FDT provides T1 service to several > smaller ISP's. Do we allocate space for them using the 25/50 rule or the > slow-start rule? These smaller ISP's that feed from us have customers > using multiple IP's (selling web space to others). Do they [the smaller > ISPs] assign space to their customers using 25/50 or slow-start? > > It seems the rules are either too simple for the real world or were > written with the idea that there are only providers and end users, and > that nobody would be both. > > ------------------------------------------------------------------ > Jon Lewis | Unsolicited commercial e-mail will > Network Administrator | be proof-read for $199/message. > Florida Digital Turnpike | > ______http://inorganic5.fdt.net/~jlewis/pgp for PGP public key____ > ---------------End of Original Message----------------- -- From: Joseph T. Klein, Titania Corporation http://www.titania.net mailto:jtk at titania.net Sent: 02:53:06 CST/CDT 09/10/97 If the Internet stumbles, it will not be because we lack for technology, vision, or motivation. It will be because we cannot set a direction and march collectively into the future. -- http://info.isoc.org/internet-history/#Future From jwkckid1 at ix.netcom.com Wed Sep 10 03:35:54 1997 From: jwkckid1 at ix.netcom.com (Jeff Williams) Date: Wed, 10 Sep 1997 08:35:54 +0100 Subject: too many routes References: Message-ID: <34164DDA.6DEF@ix.netcom.com> Joseph, Joseph T. Klein wrote: > > A system must be established that will NOT keep new entities from entering the market. > I would rather see a world with 30 or 40 backbones, vs. a world with one. Yep! I have been saying this for some time now (2 or 3 years). > > The concept of upstream providers providing address space does not work for > organizations seeking to establish national or international networks. It sure doesn't. And is the fundamental problem with the "Upstream Provider" system along with the RFC's that go along with this method (Rfc2050, Rfc 1918, RFC1917). > > If I am to build a network to 5 major meet points and contract with UUNET, and > MCI for transit, why should I put up with the InetrNIC or ARIN telling me > that I must get a block of addresses from an upstream??! How the do you expect > to establish new backbones without a loophole for new providers to get a /19 > and an ASN? Even to anty up to do this, we are in the 1 Million $$$ range. YEs! I have been trying to get this across for some time now. Glad you are on board with common sense. > > Come on registry people ... give the open market model a chance. Don't count on it. We shall have to build our own I fear. > > --- On Tue, 9 Sep 1997 23:41:50 -0400 (EDT) Jon Lewis wrote: - snip - > -- > From: Joseph T. Klein, Titania Corporation http://www.titania.net > mailto:jtk at titania.net Sent: 02:53:06 CST/CDT 09/10/97 > > If the Internet stumbles, it will not be because we lack for technology, > vision, or motivation. It will be because we cannot set a direction > and march collectively into the future. > -- http://info.isoc.org/internet-history/#Future Regards, -- Jeffrey A. Williams DIR. Internet Network Eng/SR. Java Development Eng. Information Eng. Group. IEG. INC. Phone :913-294-2375 (v-office) E-Mail jwkckid1 at ix.netcom.com From jim at ssmc.demon.co.uk Wed Sep 10 11:11:39 1997 From: jim at ssmc.demon.co.uk (Jim Rosser) Date: Wed, 10 Sep 1997 16:11:39 +0100 Subject: too many routes Message-ID: <873904263.1528939.0@ssmc.demon.co.uk> Hello! Why are you sending me this crap? Jim Rosser, jim at ssmc.demon.co.uk ---------- > From: Jeff Williams > To: Joseph T. Klein > Cc: Jon Lewis ; Michael Dillon ; naipr at arin.net > Subject: Re: too many routes > Date: Wednesday, September 10, 1997 08:35 > > Joseph, > > Joseph T. Klein wrote: > > > > A system must be established that will NOT keep new entities from entering the market. > > I would rather see a world with 30 or 40 backbones, vs. a world with one. > > Yep! I have been saying this for some time now (2 or 3 years). > > > > The concept of upstream providers providing address space does not work for > > organizations seeking to establish national or international networks. > > It sure doesn't. And is the fundamental problem with the "Upstream > Provider" system along with the RFC's that go along with this > method (Rfc2050, Rfc 1918, RFC1917). > > > > If I am to build a network to 5 major meet points and contract with UUNET, and > > MCI for transit, why should I put up with the InetrNIC or ARIN telling me > > that I must get a block of addresses from an upstream??! How the do you expect > > to establish new backbones without a loophole for new providers to get a /19 > > and an ASN? Even to anty up to do this, we are in the 1 Million $$$ range. > > YEs! I have been trying to get this across for some time now. Glad > you are on board with common sense. > > > > Come on registry people ... give the open market model a chance. > > Don't count on it. We shall have to build our own I fear. > > > > --- On Tue, 9 Sep 1997 23:41:50 -0400 (EDT) Jon Lewis wrote: > - snip - > > -- > > From: Joseph T. Klein, Titania Corporation http://www.titania.net > > mailto:jtk at titania.net Sent: 02:53:06 CST/CDT 09/10/97 > > > > If the Internet stumbles, it will not be because we lack for technology, > > vision, or motivation. It will be because we cannot set a direction > > and march collectively into the future. > > -- http://info.isoc.org/internet-history/#Future > > Regards, > -- > Jeffrey A. Williams > DIR. Internet Network Eng/SR. Java Development Eng. > Information Eng. Group. IEG. INC. > Phone :913-294-2375 (v-office) > E-Mail jwkckid1 at ix.netcom.com From Valdis.Kletnieks at VT.EDU Wed Sep 10 15:29:24 1997 From: Valdis.Kletnieks at VT.EDU (Valdis.Kletnieks at VT.EDU) Date: Wed, 10 Sep 1997 15:29:24 -0400 Subject: too many routes In-Reply-To: Your message of "Tue, 09 Sep 1997 23:41:50 EDT." References: Message-ID: <199709101929.PAA19100@black-ice.cc.vt.edu> On Tue, 09 Sep 1997 23:41:50 EDT, Jon Lewis said: > How do the rules of rfc2050 apply though when you have a hierarchy of > providers and customers, where often the customer is a provider? i.e. > picture the food chain I'm part of. UUNet provides a T1 and 2 /20 blocks > to FDT (an ISP). Should UUNet give FDT address space based on the 25%/50% > rule or the "slow-start" procedure? FDT provides T1 service to several That's for UUNet and for you to decide. Remember, the *REGISTRY* only did *one* allocation - to UUNet. It does so based on RFC2050 (plus whatever other policies they create above and beyond that). What UUNet and FDT do is market-driven. It is *intended* that they sub-delegate address space to their customers, possibly repeatedly (as in your case). On Wed, 10 Sep 97 02:53:05, Joseph T. Klein said: > If I am to build a network to 5 major meet points and contract with > UUNET, and MCI for transit, why should I put up with the InetrNIC or > ARIN telling me that I must get a block of addresses from an > upstream??! How the do you expect to establish new backbones without > a loophole for new providers to get a /19 >and an ASN? Even to anty > up to do this, we are in the 1 Million $$$ range. To run a sucessful startup, you gotta be capitalized. Life sucks. I am sure that *IF YOU COME IN WITH A BUSINESS PLAN*, including capital sources, projected hookups, and the like, the guys at ARIN will be willing to listen. But seriously - anybody who's shelling out for interconnects at 5 different meet points had *BETTER* have enough business lined up to fill a /19, or you're going to go *broke* shelling out to UUNET and MCI for the lines. A /19 is only 8K addresses. If you don't have a business plan to get enough to fill a /19 (and due to subnetting etc, you may be able to get close to full on 4K actual allocated or even less), not getting a /19 from ARIN is the least of your worries. Cash Flow is probably the biggest. -- Valdis Kletnieks Computer Systems Senior Engineer Virginia Tech -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 284 bytes Desc: not available URL: From handley at admin.microserve.net Wed Sep 10 15:46:07 1997 From: handley at admin.microserve.net (Mike Handley) Date: Wed, 10 Sep 1997 15:46:07 -0400 (EDT) Subject: too many routes In-Reply-To: <873904263.1528939.0@ssmc.demon.co.uk> Message-ID: I have also gotten several! Why? Mike Handley handley at microserve.net On Wed, 10 Sep 1997, Jim Rosser wrote: > Hello! > > Why are you sending me this crap? > > Jim Rosser, jim at ssmc.demon.co.uk > > ---------- > > From: Jeff Williams > > To: Joseph T. Klein > > Cc: Jon Lewis ; Michael Dillon > ; naipr at arin.net > > Subject: Re: too many routes > > Date: Wednesday, September 10, 1997 08:35 > > > > Joseph, > > > > Joseph T. Klein wrote: > > > > > > A system must be established that will NOT keep new entities from > entering the market. > > > I would rather see a world with 30 or 40 backbones, vs. a world with > one. > > > > Yep! I have been saying this for some time now (2 or 3 years). > > > > > > The concept of upstream providers providing address space does not > work for > > > organizations seeking to establish national or international networks. > > > > It sure doesn't. And is the fundamental problem with the "Upstream > > Provider" system along with the RFC's that go along with this > > method (Rfc2050, Rfc 1918, RFC1917). > > > > > > If I am to build a network to 5 major meet points and contract with > UUNET, and > > > MCI for transit, why should I put up with the InetrNIC or ARIN telling > me > > > that I must get a block of addresses from an upstream??! How the do you > expect > > > to establish new backbones without a loophole for new providers to get > a /19 > > > and an ASN? Even to anty up to do this, we are in the 1 Million $$$ > range. > > > > YEs! I have been trying to get this across for some time now. Glad > > you are on board with common sense. > > > > > > Come on registry people ... give the open market model a chance. > > > > Don't count on it. We shall have to build our own I fear. > > > > > > --- On Tue, 9 Sep 1997 23:41:50 -0400 (EDT) Jon Lewis > wrote: > > - snip - > > > -- > > > From: Joseph T. Klein, Titania Corporation http://www.titania.net > > > mailto:jtk at titania.net Sent: 02:53:06 CST/CDT > 09/10/97 > > > > > > If the Internet stumbles, it will not be because we lack for > technology, > > > vision, or motivation. It will be because we cannot set a direction > > > and march collectively into the future. > > > -- http://info.isoc.org/internet-history/#Future > > > > Regards, > > -- > > Jeffrey A. Williams > > DIR. Internet Network Eng/SR. Java Development Eng. > > Information Eng. Group. IEG. INC. > > Phone :913-294-2375 (v-office) > > E-Mail jwkckid1 at ix.netcom.com > ~ ---------------------------------------------------------------------------- Mike Handley MicroServe Information Systems Mail: handley at admin.microserve.net Engineering Tip of the Week/Month by Mike Handley: ---If you could unknow anyone you already know, whom would it be?---- ------------------------------------------------------------------------------ From michael at priori.net Wed Sep 10 17:04:14 1997 From: michael at priori.net (Michael Dillon) Date: Wed, 10 Sep 1997 14:04:14 -0700 Subject: Explanation of why you are getting this mail In-Reply-To: References: <873904263.1528939.0@ssmc.demon.co.uk> Message-ID: >I have also gotten several! Why? > >> Why are you sending me this crap? If anyone else on this mailing list does not wish to receive future mail about the American Registry for Internet Numbers (ARIN) then follow these instructions: Send an email to naipr-request at arin.net In the body of the message, NOT the subject, place the single word unsubscribe That's it! You should receive a response from the mailserver indicating that you have been dropped from the list. If the response indicates that your address is not currently on the list then it's time to start asking around your company and find which one of your co-workers is forwarding you this mail. ******************************************************** Michael Dillon voice: +1-650-482-2840 Senior Systems Architect fax: +1-650-482-2844 PRIORI NETWORKS, INC. http://www.priori.net "The People You Know. The People You Trust." ******************************************************** From jwkckid1 at ix.netcom.com Wed Sep 10 13:51:57 1997 From: jwkckid1 at ix.netcom.com (Jeff Williams) Date: Wed, 10 Sep 1997 18:51:57 +0100 Subject: too many routes References: <199709101929.PAA19100@black-ice.cc.vt.edu> Message-ID: <3416DE3D.1FAB@ix.netcom.com> Valdis and all Valdis.Kletnieks at VT.EDU wrote: > > On Tue, 09 Sep 1997 23:41:50 EDT, Jon Lewis said: > > How do the rules of rfc2050 apply though when you have a hierarchy of > > providers and customers, where often the customer is a provider? i.e. > > picture the food chain I'm part of. UUNet provides a T1 and 2 /20 blocks > > to FDT (an ISP). Should UUNet give FDT address space based on the 25%/50% > > rule or the "slow-start" procedure? FDT provides T1 service to several > > That's for UUNet and for you to decide. Remember, the *REGISTRY* only > did *one* allocation - to UUNet. It does so based on RFC2050 (plus > whatever other policies they create above and beyond that). What UUNet > and FDT do is market-driven. It is *intended* that they sub-delegate > address space to their customers, possibly repeatedly (as in your case). Where on ARIN's site is there a clear and percise policy stating that RFC2050 can be suplemented as per your statment, "(plus whatever other policies they create above and beyond that)" ? In section 2 of RFC2050 a buisness plane is not mentioned directly as a qualification for any allocation. This effects new ISP's especially regional ones and smaller. > > On Wed, 10 Sep 97 02:53:05, Joseph T. Klein said: > > > If I am to build a network to 5 major meet points and contract with > > UUNET, and MCI for transit, why should I put up with the InetrNIC or > > ARIN telling me that I must get a block of addresses from an > > upstream??! How the do you expect to establish new backbones without > > a loophole for new providers to get a /19 >and an ASN? Even to anty > > up to do this, we are in the 1 Million $$$ range. > > To run a sucessful startup, you gotta be capitalized. Life sucks. I am > sure that *IF YOU COME IN WITH A BUSINESS PLAN*, including capital sources, > projected hookups, and the like, the guys at ARIN will be willing to listen. > But seriously - anybody who's shelling out for interconnects at 5 different > meet points had *BETTER* have enough business lined up to fill a /19, > or you're going to go *broke* shelling out to UUNET and MCI for the lines. Hell we are capitolized to the tune of $12m and yet I have been told on this list that that will not have any effect on weather I can recieve a /19 or not. That RFC2050 will be the basis on any allocation. So I have no idea where you get this information. If the ARIN guys are willing to listen it is news to me. Kim Hubbard, can you comment here? And really this is not a sound basis for determining allocations. There should either be a clear and percise policy backed by a modification to RFC2050 amended with such language that you mention or we are just playing little favorites game here. That is not good buisness nor good for smaller communities which wish to have their own ISP's. > > A /19 is only 8K addresses. If you don't have a business plan to get > enough to fill a /19 (and due to subnetting etc, you may be able to > get close to full on 4K actual allocated or even less), not getting a /19 > from ARIN is the least of your worries. Cash Flow is probably the biggest. This would usually be the case, I must admit. But not in all cases, but I would agree most. However that is really besides th point really. (See above my comments). > > -- > Valdis Kletnieks > Computer Systems Senior Engineer > Virginia Tech > > --------------------------------------------------------------- > > Part 1.2 Type: application/pgp-signature Regards, -- Jeffrey A. Williams DIR. Internet Network Eng/SR. Java Development Eng. Information Eng. Group. IEG. INC. Phone :913-294-2375 (v-office) E-Mail jwkckid1 at ix.netcom.com From jlewis at inorganic5.fdt.net Sun Sep 14 13:58:08 1997 From: jlewis at inorganic5.fdt.net (Jon Lewis) Date: Sun, 14 Sep 1997 13:58:08 -0400 (EDT) Subject: too many routes In-Reply-To: Message-ID: On Tue, 9 Sep 1997, Jon Lewis wrote: > How do the rules of rfc2050 apply though when you have a hierarchy of > providers and customers, where often the customer is a provider? i.e. > picture the food chain I'm part of. UUNet provides a T1 and 2 /20 blocks > to FDT (an ISP). Should UUNet give FDT address space based on the 25%/50% > rule or the "slow-start" procedure? FDT provides T1 service to several > smaller ISP's. Do we allocate space for them using the 25/50 rule or the > slow-start rule? These smaller ISP's that feed from us have customers > using multiple IP's (selling web space to others). Do they [the smaller > ISPs] assign space to their customers using 25/50 or slow-start? I don't think I ever saw a reply to this, and the answers (if there are official ones) to the above questions are likely to be of considerable importance to me very soon. Are there answers, or is rfc2050 open to interpretation by the top level providers...or is it not even relevant to their dealings with their customers? ------------------------------------------------------------------ Jon Lewis | Unsolicited commercial e-mail will Network Administrator | be proof-read for $199/message. Florida Digital Turnpike | ______http://inorganic5.fdt.net/~jlewis/pgp for PGP public key____ From jwkckid1 at ix.netcom.com Sun Sep 14 20:50:12 1997 From: jwkckid1 at ix.netcom.com (Jeff Williams) Date: Mon, 15 Sep 1997 01:50:12 +0100 Subject: too many routes References: Message-ID: <341C8644.52A5@ix.netcom.com> Jon, Jon Lewis wrote: > > On Tue, 9 Sep 1997, Jon Lewis wrote: > > > How do the rules of rfc2050 apply though when you have a hierarchy of > > providers and customers, where often the customer is a provider? i.e. > > picture the food chain I'm part of. UUNet provides a T1 and 2 /20 blocks > > to FDT (an ISP). Should UUNet give FDT address space based on the 25%/50% > > rule or the "slow-start" procedure? FDT provides T1 service to several > > smaller ISP's. Do we allocate space for them using the 25/50 rule or the > > slow-start rule? These smaller ISP's that feed from us have customers > > using multiple IP's (selling web space to others). Do they [the smaller > > ISPs] assign space to their customers using 25/50 or slow-start? > > I don't think I ever saw a reply to this, and the answers (if there are > official ones) to the above questions are likely to be of considerable > importance to me very soon. > > Are there answers, or is rfc2050 open to interpretation by the top level > providers...or is it not even relevant to their dealings with their > customers? I don't know myself. I have been asking this question for several months now and no "Official" answer has been forthcomming. > > ------------------------------------------------------------------ > Jon Lewis | Unsolicited commercial e-mail will > Network Administrator | be proof-read for $199/message. > Florida Digital Turnpike | > ______http://inorganic5.fdt.net/~jlewis/pgp for PGP public key____ Regards, -- Jeffrey A. Williams DIR. Internet Network Eng/SR. Java Development Eng. Information Eng. Group. IEG. INC. Phone :913-294-2375 (v-office) E-Mail jwkckid1 at ix.netcom.com From kimh at internic.net Wed Sep 17 14:18:37 1997 From: kimh at internic.net (Kim Hubbard) Date: Wed, 17 Sep 1997 14:18:37 -0400 (EDT) Subject: too many routes In-Reply-To: from "Jon Lewis" at Sep 14, 97 01:58:08 pm Message-ID: <199709171818.OAA13719@ops.internic.net> > UUNET would allocate space to their downstream ISPs using the same criteria listed in the allocation portion of RFC2050. ISP customers are treated as ISPs - not end-users, meaning their 1 year requirement is not taken into consideration. And I know some of you may be thinking that this isn't right, but *most* ISPs do not know what their network or customer base is going to look like in a year (they know what they want it to look like) so it's better to base the allocation size on three months. Kim > On Tue, 9 Sep 1997, Jon Lewis wrote: > > > How do the rules of rfc2050 apply though when you have a hierarchy of > > providers and customers, where often the customer is a provider? i.e. > > picture the food chain I'm part of. UUNet provides a T1 and 2 /20 blocks > > to FDT (an ISP). Should UUNet give FDT address space based on the 25%/50% > > rule or the "slow-start" procedure? FDT provides T1 service to several > > smaller ISP's. Do we allocate space for them using the 25/50 rule or the > > slow-start rule? These smaller ISP's that feed from us have customers > > using multiple IP's (selling web space to others). Do they [the smaller > > ISPs] assign space to their customers using 25/50 or slow-start? > > I don't think I ever saw a reply to this, and the answers (if there are > official ones) to the above questions are likely to be of considerable > importance to me very soon. > > Are there answers, or is rfc2050 open to interpretation by the top level > providers...or is it not even relevant to their dealings with their > customers? > > ------------------------------------------------------------------ > Jon Lewis | Unsolicited commercial e-mail will > Network Administrator | be proof-read for $199/message. > Florida Digital Turnpike | > ______http://inorganic5.fdt.net/~jlewis/pgp for PGP public key____ >