Re: [IPsec] Traffic Visibility Future

2010-01-08 Thread Bhatia, Manav (Manav)
> > > > In an ideal world maybe. Sometimes the netwwork needs to > mark traffic > > at the edge switch but doesn't 'trust' the endpoint to do it. So > > typically it would be done with policy rules. > > We're talking about making changes to the end nodes anyways, > so why not > make them handle

Re: [IPsec] Traffic Visibility Future

2010-01-08 Thread Nicolas Williams
On Fri, Jan 08, 2010 at 11:14:06AM -0800, Joseph Tardo wrote: > -Original Message- > > Also, anything QoS related should really happen outside ESP anyways. > > > In an ideal world maybe. Sometimes the netwwork needs to mark traffic > at the edge switch but doesn't 'trust' the endpoint

Re: [IPsec] Traffic Visibility Future

2010-01-08 Thread Joseph Tardo
-Original Message- Also, anything QoS related should really happen outside ESP anyways. Nico -- In an ideal world maybe. Sometimes the netwwork needs to mark traffic at the edge switch but doesn't 'trust' the endpoint to do it. So typically it would be done with policy rules. Thank

Re: [IPsec] Traffic Visibility Future

2010-01-08 Thread Bhatia, Manav (Manav)
2 AM > To: Bhatia, Manav (Manav) > Cc: Dan Harkins; Jack Kohn; ipsec@ietf.org > Subject: Re: [IPsec] Traffic Visibility Future > > > I would be very much interested in a real-world example instead > of hypotheticals. What service provider is asking for WESP? > > Dan.

Re: [IPsec] Traffic Visibility Future

2010-01-07 Thread Nicolas Williams
On Fri, Jan 08, 2010 at 07:39:47AM +0530, Jack Kohn wrote: > > Yes, but it's useful to know what they could possibly do.  If a > > middlebox wants to drop all encrypted IPsec traffic then it needs to > > know if it's encrypted.  Knowing WESP -> ESP-null, ESP -> unknown, is > > sufficient. > > Yup,

Re: [IPsec] Traffic Visibility Future

2010-01-07 Thread Dan Harkins
I would be very much interested in a real-world example instead of hypotheticals. What service provider is asking for WESP? Dan. On Thu, January 7, 2010 6:15 pm, Bhatia, Manav (Manav) wrote: > Dan, > > [clipped] > >> Because it's unnecessary bloat that another group may not have any >> use

Re: [IPsec] Traffic Visibility Future

2010-01-07 Thread Bhatia, Manav (Manav)
Dan, [clipped] > Because it's unnecessary bloat that another group may not have any > use for. ESP-null could be used, for instance, to protect > packets in an > EGP routing protocol. There is no need for WESP in such an > environment. EGP routing protocols, by definition and design, will t

Re: [IPsec] Traffic Visibility Future

2010-01-07 Thread Jack Kohn
> > Yes, but it's useful to know what they could possibly do.  If a > middlebox wants to drop all encrypted IPsec traffic then it needs to > know if it's encrypted.  Knowing WESP -> ESP-null, ESP -> unknown, is > sufficient. Yup, and this is what i was alluding to. All middle boxes may not want to

Re: [IPsec] Traffic Visibility Future

2010-01-07 Thread Nicolas Williams
On Thu, Jan 07, 2010 at 03:51:04PM -0900, Melinda Shore wrote: > On Jan 7, 2010, at 3:45 PM, Jack Kohn wrote: > >I am just trying to understand what a WESP powered middle box thats > >interested in deep inspecting packets, should do when it sees a native > >ESP packet. Should it make an attempt to

Re: [IPsec] Traffic Visibility Future

2010-01-07 Thread Dan Harkins
On Thu, January 7, 2010 4:39 pm, Jack Kohn wrote: > On Fri, Jan 8, 2010 at 5:51 AM, Dan Harkins wrote: >> >>  Hi Jack, >> >> On Thu, January 7, 2010 4:06 pm, Jack Kohn wrote: >>> Folks, >>> >>> Some questions. >>> >>> o In a steady state, where we are using WESP only for ESP-NULL, what >>> should

Re: [IPsec] Traffic Visibility Future

2010-01-07 Thread Melinda Shore
On Jan 7, 2010, at 3:45 PM, Jack Kohn wrote: I am just trying to understand what a WESP powered middle box thats interested in deep inspecting packets, should do when it sees a native ESP packet. Should it make an attempt to parse it based on heuristics (which i completely resent) or should it tr

Re: [IPsec] Traffic Visibility Future

2010-01-07 Thread Jack Kohn
On Fri, Jan 8, 2010 at 5:55 AM, Melinda Shore wrote: > On Jan 7, 2010, at 3:06 PM, Jack Kohn wrote: >> >> o In a steady state, where we are using WESP only for ESP-NULL, what >> should a middle box do when it sees  ESP traffic, besides >> hyperventilating and throwing up? > > How would that inform

Re: [IPsec] Traffic Visibility Future

2010-01-07 Thread Jack Kohn
On Fri, Jan 8, 2010 at 5:51 AM, Dan Harkins wrote: > >  Hi Jack, > > On Thu, January 7, 2010 4:06 pm, Jack Kohn wrote: >> Folks, >> >> Some questions. >> >> o In a steady state, where we are using WESP only for ESP-NULL, what >> should a middle box do when it sees  ESP traffic, besides >> hyperven

Re: [IPsec] Traffic Visibility Future

2010-01-07 Thread Melinda Shore
On Jan 7, 2010, at 3:06 PM, Jack Kohn wrote: o In a steady state, where we are using WESP only for ESP-NULL, what should a middle box do when it sees ESP traffic, besides hyperventilating and throwing up? How would that information be used here? Do you want to specify middlebox behavior? In

Re: [IPsec] Traffic Visibility Future

2010-01-07 Thread Dan Harkins
Hi Jack, On Thu, January 7, 2010 4:06 pm, Jack Kohn wrote: > Folks, > > Some questions. > > o In a steady state, where we are using WESP only for ESP-NULL, what > should a middle box do when it sees ESP traffic, besides > hyperventilating and throwing up? Should it run heuristics (dang! no) >

[IPsec] Traffic Visibility Future

2010-01-07 Thread Jack Kohn
Folks, Some questions. o In a steady state, where we are using WESP only for ESP-NULL, what should a middle box do when it sees ESP traffic, besides hyperventilating and throwing up? Should it run heuristics (dang! no) or should it simply assume that the packet is encrypted and do whatever the l