[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: geo short vs long term? [Re: Geo pros and cons]
I would, without objection, like to wholly withdraw the entire
mail message that I sent and refer everyone to Noel's mail.
Please excuse my temporary insanity/stupidity. I know better.
Tony
| -----Original Message-----
| From: Christian Huitema [mailto:huitema@windows.microsoft.com]
| Sent: Thursday, April 03, 2003 3:33 PM
| To: Tony Li; J. Noel Chiappa; multi6@ops.ietf.org
| Subject: RE: geo short vs long term? [Re: Geo pros and cons]
|
|
| Tony,
|
| There are a few issues with your maths.
|
| First, the 100% per year figure relates to the traffic
| growth, not the
| growth in the number of hosts. The number of hosts does
| grow, but not
| nearly as fast. The host count measured by the Internet
| domain survey
| grew 20% last year, but that does not account for hosts
| behind NAT. If
| you assume that the traffic growth is caused equally by
| larger pipes and
| by more devices, then you get a growth in the number of
| hosts of about
| 40% per year.
|
| Second, the number of sites does not grow at the same pace
| as the number
| of hosts. For example, the number of PC per household or
| per office is
| increasing. The hypothetical 40% host grow could be split between a
| growth in the number of connected sites (new homes, new
| offices) and a
| growth in the number of hosts per site. This would result in a 20%
| yearly growth for the number of sites.
|
| Most of the sites being added are small sites, which are the least
| likely to be multihomed -- the big sites are already
| connected, probably
| already multihomed. So, there is a strong case that the
| 10% figure is
| probably an overstatement. But let's keep it.
|
| The cost of maintaining the routing tables is somewhere
| between O(N.log
| N) and O(N^2). Given the current size of 100,000 entries,
| a 20% increase
| would correspond to an increase of 22% to 44%.
|
| Moore's law corresponds to an increase of about 60%
| yearly. 60% beats
| 22%, or even 44%. There is no reason to panic.
|
| There is also no particular reason to love geographic
| addresses, but
| that is another issue.
|
| -- Christian Huitema
|
| > -----Original Message-----
| > From: Tony Li [mailto:Tony.Li@procket.com]
| > Sent: Thursday, April 03, 2003 12:10 PM
| > To: J. Noel Chiappa; multi6@ops.ietf.org
| >
| >
| >
| > Noel, well said.
| >
| > More points: there are worse things than no decision. That
| > would be the clearly wrong decision. We could, for example,
| > choose to aggregate locators based on the lexographical
| > ordering of the user's first name. We could choose not to
| > aggregate and end up with another swamp.
| >
| > As to the issues with 'proving' that geo won't work, let me
| > point out the very simple reasoning:
| >
| > - The Internet is continually growing at an exponential rate.
| > Most people seem to peg the growth rate at 100% per year
| > currently. The exact number is not an issue.
| >
| > - In the past, we've estimated that 10% of all sites would
| > multi-home. Let's assume a constant rate of 10% of the
| > world is an exception to the default aggregation rules
| > that we pick.
| >
| > - From the above two, we can reason that our exception rate
| > is going to continue to grow exponentially. Note that
| > the rate of absolute growth is more of an issue than the
| > exception rate.
| >
| > - Moore's law for memory suggests that memory sizes will
| > double about every two years. However, memory speeds will
| > not keep up.
| >
| > - Packet lookups are a function of memory bandwidth, so to
| > sustain Internet bandwidth growth of 100% per year, we need
| > to also increase memory bandwidth by about 100% per year.
| > Using bigger, slower memories is not a realistic option.
| >
| > - Thus, the routing table really needs to be constrained to
| > grow at about Moore's law for memory.
| >
| > - If the exceptions are growing at about 100% per year, and
| > the memories are growing at about 100% every TWO years, then
| > regardless of the starting point, the exceptions will overtake
| > technology.
| >
| > - Therefore, we must find some mechanism that prevents the
| > exceptions from growing at 100% per year. In short, the
| > number of longer prefixes that are injected into routing
| > cannot be a constant fraction of the number of sites that
| > join.
| >
| > - Since everyone and their brother will want an exception
| > for anything that they want to do that is outside of the
| > norm, the norm MUST support almost every possible situation.
| > Multihoming, in particular, must not cause exceptions.
| > Even a constant percentage of multihomers must not cause
| > exceptions.
| >
| > - For reasons that I've already explained, the economics
| > of links in a geo system cause many sites to be exceptions.
| >
| > - Therefore, geo addressing leads to a system that will not
| > scale for the long term.
| >
| > QED
| >
| > Tony
| >
| >
| > | -----Original Message-----
| > | From: J. Noel Chiappa [mailto:jnc@ginger.lcs.mit.edu]
| > | Sent: Thursday, April 03, 2003 4:38 AM
| > | To: multi6@ops.ietf.org
| > | Cc: jnc@ginger.lcs.mit.edu
| > | Subject: Re: geo short vs long term? [Re: Geo pros and cons]
| > |
| > |
| > | Look, everyone, this is all really stupid.
| > |
| > | Geographic addressing has been discussed extensively about
| > | 17 times in the
| > | IETF, and every time it has been rejected. Discussing it
| > | one more time is not
| > | going to change this. There is *never* going to be a rough
| > | consensus *in
| > | favour of* geographic addressing. There will *always* be a
| > | lot of people
| > | against it - enough to stop it in the proposal stage.
| > |
| > | The really sad thing is that something productive might
| > | have been done with
| > | all this time and energy that's being wasted.
| > |
| > | Noel
| > |
| > |
| >
|
|
|