next version of RIPE-501, v.2
Hi, Please find attached document with many ideas and suggestions included, LB spec is also there. As usual, all comments warmly welcome. Cheers, Jan
On 22 jul 2011, at 10:02, Jan Zorz @ go6.si wrote:
Hi,
Please find attached document with many ideas and suggestions included, LB spec is also there.
As usual, all comments warmly welcome.
Cheers, Jan <Requirements-for-IPv6-in-ICT-equipment-v.2.pdf>
As a small request, can people please post plain text only documents to this list. Makes it much easier to track comments and feedback. Thanks, MarcoH -- "Good tests kill flawed theories; we remain alive to guess again"
Hi, relating on following sentence: "If MPLS Traffic Engineering is used in combination with IS-IS routing protocol, the equipment MUST support "M-ISIS: Multi-Topology (MT) Routing in Intermediate System to Intermediate Systems (IS-ISs)" [RFC 5120]" I think the Multitopology ISIS can be used also without TE ad said in RFC 5120 with IPv6: "This document describes how to run, within a single IS-IS domain, a set of independent IP topologies that we call Multi-Topologies (MTs). This MT extension can be used for a variety of purposes, such as an in-band management network "on top" of the original IGP topology, maintaining separate IGP routing domains for isolated multicast or IPv6 islands within the backbone, or forcing a subset of an address space to follow a different topology. ".... Jan Zorz @ go6.si ha scritto:
Hi,
Please find attached document with many ideas and suggestions included, LB spec is also there.
As usual, all comments warmly welcome.
Cheers, Jan
-- Isacco Fontana Trentino Network s.r.l. A socio Unico Direzione Servizi Responsabile Area Ingegneria di Rete Via Gilli, 2 - 38100 TRENTO Tel (+39) 0461.020200 Fax (+39) 0461.020201 http://as12835.peeringdb.com/ -------------------------------------------------------------------------------------------------------------------------------- Cap. Soc. sottoscritto € 18.213.248,00. i.v. - REG. IMP. C.F. e P. IVA 01904880224 E-mail: sede@trentinonetwork.it Società soggetta a direzione e controllo da parte della Provincia Autonoma di Trento. C.F. e P. IVA 00337460224 Questo messaggio è indirizzato esclusivamente ai destinatari in intestazione, può contenere informazioni protette e riservate ai sensi della normativa vigente e ne è vietato qualsiasi impiego diverso da quello per cui è stato inviato. Se lo avete ricevuto per errore siete pregati di eliminarlo in ogni sua parte e di avvisare il mittente. --------------------------------------------------------------------------------------------------------------------------------
On 7/22/11 3:57 PM, Isacco Fontana wrote:
Hi, relating on following sentence: "If MPLS Traffic Engineering is used in combination with IS-IS routing protocol, the equipment MUST support "M-ISIS: Multi-Topology (MT) Routing in Intermediate System to Intermediate Systems (IS-ISs)" [RFC 5120]"
I think the Multitopology ISIS can be used also without TE ad said in RFC 5120 with IPv6: "This document describes how to run, within a single IS-IS domain, a set of independent IP topologies that we call Multi-Topologies (MTs). This MT extension can be used for a variety of purposes, such as an in-band management network "on top" of the original IGP topology, maintaining separate IGP routing domains for isolated multicast or IPv6 islands within the backbone, or forcing a subset of an address space to follow a different topology. "....
thnx for comment. What would be suggested change in text? Let's see if Ivan "mpls-master" Pepelnjak agree on proposed change :) Cheers, Jan
IS-IS MT is highly desirable in most circumstances anyway, but we haven't considered that a good-enough reason to make it MANDATORY. However, if you run MPLS TE without MT, you get black hole routing the moment the first autoroute MPLS TE tunnel is established; thus we've made IS-IS MT MANDATORY for networks running MPLS TE. Details here: http://blog.ioshints.info/2010/03/is-ismpls-tenative-ipv6fail.html However, I'm perfectly happy if the WG decides to make IS-IS MT mandatory in all cases (would make sense anyway). Cheers, Ivan
-----Original Message----- From: ipv6-wg-admin@ripe.net [mailto:ipv6-wg-admin@ripe.net] On Behalf Of Jan Zorz @ go6.si Sent: Saturday, July 23, 2011 12:12 AM To: ipv6-wg@ripe.net Subject: Re: [ipv6-wg] next version of RIPE-501, v.2
On 7/22/11 3:57 PM, Isacco Fontana wrote:
Hi, relating on following sentence: "If MPLS Traffic Engineering is used in combination with IS-IS routing protocol, the equipment MUST support "M-ISIS: Multi-Topology (MT) Routing in Intermediate System to Intermediate Systems (IS-ISs)" [RFC 5120]"
I think the Multitopology ISIS can be used also without TE ad said in RFC 5120 with IPv6: "This document describes how to run, within a single IS-IS domain, a set of independent IP topologies that we call Multi-Topologies (MTs). This MT extension can be used for a variety of purposes, such as an in-band management network "on top" of the original IGP topology, maintaining separate IGP routing domains for isolated multicast or IPv6 islands within the backbone, or forcing a subset of an address space to follow a different topology. "....
thnx for comment.
What would be suggested change in text?
Let's see if Ivan "mpls-master" Pepelnjak agree on proposed change :)
Cheers, Jan
On 7/23/11 9:37 AM, Ivan Pepelnjak wrote:
IS-IS MT is highly desirable in most circumstances anyway, but we haven't considered that a good-enough reason to make it MANDATORY.
However, if you run MPLS TE without MT, you get black hole routing the moment the first autoroute MPLS TE tunnel is established; thus we've made IS-IS MT MANDATORY for networks running MPLS TE.
Details here: http://blog.ioshints.info/2010/03/is-ismpls-tenative-ipv6fail.html
However, I'm perfectly happy if the WG decides to make IS-IS MT mandatory in all cases (would make sense anyway).
Currently, this is mandatory only "If MPLS Traffic Engineering is used in combination with IS-IS routing protocol" What percentage of equipment we exclude as routers if we make this unconditionally mandatory? I know for at least Mikrotik routers are excluded, as they do not support IS-IS at all (and they are quite used in small/medium companies environment). Opinions? Cheers, Jan
If the equipment doesn't support IS-IS, then you can't expect it to support IS-IS MT ;) Anyway, making IS-IS mandatory makes absolutely no sense; it's rarely used in Enterprise environments. What we could do is to change the current requirement into "If the equipment supports IS-IS routing protocol, it MUST support IS-IS MT ..." (don’t cut/paste, use the wording from the document ;) Ivan
Currently, this is mandatory only "If MPLS Traffic Engineering is used in combination with IS-IS routing protocol"
What percentage of equipment we exclude as routers if we make this unconditionally mandatory?
I know for at least Mikrotik routers are excluded, as they do not support IS-IS at all (and they are quite used in small/medium companies environment).
Opinions?
Cheers, Jan
Hi,
If the equipment doesn't support IS-IS, then you can't expect it to support IS-IS MT ;)
Anyway, making IS-IS mandatory makes absolutely no sense; it's rarely used in Enterprise environments.
What we could do is to change the current requirement into "If the equipment supports IS-IS routing protocol, it MUST support IS-IS MT ..." (don’t cut/paste, use the wording from the document ;)
Sounds like good advise :) Thanks, Sander
I agree ... Isacco Inviato da iPad Il giorno 24/lug/2011, alle ore 16:12, Sander Steffann <sander@steffann.nl> ha scritto:
Hi,
If the equipment doesn't support IS-IS, then you can't expect it to support IS-IS MT ;)
Anyway, making IS-IS mandatory makes absolutely no sense; it's rarely used in Enterprise environments.
What we could do is to change the current requirement into "If the equipment supports IS-IS routing protocol, it MUST support IS-IS MT ..." (don’t cut/paste, use the wording from the document ;)
Sounds like good advise :)
Thanks,
Sander
On 7/24/11 12:32 PM, Ivan Pepelnjak wrote:
If the equipment doesn't support IS-IS, then you can't expect it to support IS-IS MT ;)
;)
Anyway, making IS-IS mandatory makes absolutely no sense; it's rarely used in Enterprise environments.
What we could do is to change the current requirement into "If the equipment supports IS-IS routing protocol, it MUST support IS-IS MT ..." (don’t cut/paste, use the wording from the document ;)
hmm... I think we should not say "If equipment supports", as *we* are writing document defining what equipment must and should support. ;) Probably it would be a good thing to say something like: "If IS-IS [RFC5308] is requested, then IS-IS MT must be supported" or maybe we just move from optional to mandatory this section: "When IS-IS routing protocol is requested, the equipment SHOULD support "M-ISIS: Multi-Topology (MT) Routing in Intermediate System to Intermediate Systems (IS-ISs)" [RFC 5120] (highly recommended)" ...change the first word "When" to "If" and remove "(highly recommended)" :) What do you guys think? /jan
Hi,
"When IS-IS routing protocol is requested, the equipment SHOULD support "M-ISIS: Multi-Topology (MT) Routing in Intermediate System to Intermediate Systems (IS-ISs)" [RFC 5120] (highly recommended)" What do you guys think?
I would put this in 'mandatory': - If the IS-IS routing protocol is requested the equipment MUST support "M-ISIS: Multi-Topology (MT) Routing in Intermediate System to Intermediate Systems (IS-ISs)" [RFC 5120] Met vriendelijke groet, Sander Steffann
On 7/24/11 10:03 PM, Sander Steffann wrote:
I would put this in 'mandatory':
- If the IS-IS routing protocol is requested the equipment MUST support "M-ISIS: Multi-Topology (MT) Routing in Intermediate System to Intermediate Systems (IS-ISs)" [RFC 5120]
ok, I see you already updated the draft on Gdocs. One question, should we use "MUST" with capital letters or we use normal "must" in this "if" sentences? Currently we use mix :) Cheers, Jan
Hi,
I would put this in 'mandatory':
- If the IS-IS routing protocol is requested the equipment MUST support "M-ISIS: Multi-Topology (MT) Routing in Intermediate System to Intermediate Systems (IS-ISs)" [RFC 5120]
ok, I see you already updated the draft on Gdocs.
I didn't put this change in yet though :)
One question, should we use "MUST" with capital letters or we use normal "must" in this "if" sentences?
We are not writing an RFC so I think we should use lowercase. Met vriendelijke groet, Sander Steffann
On 7/25/11 1:21 AM, Sander Steffann wrote:
Hi,
I would put this in 'mandatory':
- If the IS-IS routing protocol is requested the equipment MUST support "M-ISIS: Multi-Topology (MT) Routing in Intermediate System to Intermediate Systems (IS-ISs)" [RFC 5120]
ok, I see you already updated the draft on Gdocs.
I didn't put this change in yet though :)
ah, ok, my mistake. Done, changed.
One question, should we use "MUST" with capital letters or we use normal "must" in this "if" sentences?
We are not writing an RFC so I think we should use lowercase.
corrected. Cheers, Jan
I agree. I think the use of MT-ISIS when we have ipv6 is always desirable not only when we have TE. Isacco Inviato da iPad Il giorno 23/lug/2011, alle ore 09:37, "Ivan Pepelnjak" <ip@ioshints.info> ha scritto:
IS-IS MT is highly desirable in most circumstances anyway, but we haven't considered that a good-enough reason to make it MANDATORY.
However, if you run MPLS TE without MT, you get black hole routing the moment the first autoroute MPLS TE tunnel is established; thus we've made IS-IS MT MANDATORY for networks running MPLS TE.
Details here: http://blog.ioshints.info/2010/03/is-ismpls-tenative-ipv6fail.html
However, I'm perfectly happy if the WG decides to make IS-IS MT mandatory in all cases (would make sense anyway).
Cheers, Ivan
-----Original Message----- From: ipv6-wg-admin@ripe.net [mailto:ipv6-wg-admin@ripe.net] On Behalf Of Jan Zorz @ go6.si Sent: Saturday, July 23, 2011 12:12 AM To: ipv6-wg@ripe.net Subject: Re: [ipv6-wg] next version of RIPE-501, v.2
On 7/22/11 3:57 PM, Isacco Fontana wrote:
Hi, relating on following sentence: "If MPLS Traffic Engineering is used in combination with IS-IS routing protocol, the equipment MUST support "M-ISIS: Multi-Topology (MT) Routing in Intermediate System to Intermediate Systems (IS-ISs)" [RFC 5120]"
I think the Multitopology ISIS can be used also without TE ad said in RFC 5120 with IPv6: "This document describes how to run, within a single IS-IS domain, a set of independent IP topologies that we call Multi-Topologies (MTs). This MT extension can be used for a variety of purposes, such as an in-band management network "on top" of the original IGP topology, maintaining separate IGP routing domains for isolated multicast or IPv6 islands within the backbone, or forcing a subset of an address space to follow a different topology. "....
thnx for comment.
What would be suggested change in text?
Let's see if Ivan "mpls-master" Pepelnjak agree on proposed change :)
Cheers, Jan
participants (5)
-
Isacco Fontana -
Ivan Pepelnjak -
Jan Zorz ＠ go6.si -
Marco Hogewoning -
Sander Steffann