North American Network Operators Group

Date Prev | Date Next | Date Index | Thread Index | Author Index | Historical

Re: "Simple" Multi-Homing ? (was Re: CIDR Report)

  • From: Greg A. Woods
  • Date: Tue May 16 11:31:00 2000

[ On Tuesday, May 16, 2000 at 00:59:55 (-0400), Todd Sandor wrote: ]
> Subject: "Simple" Multi-Homing ? (was Re: CIDR Report)
>
> Our requirements are:
> - we needs link redundancy (a single location) \226 our web service requires 24 x 7
> availability.

That shouldn't be too hard to achieve if you can get Bell to co-operate.

What you might want to look at is provisioning two different services,
eg. the T1 you have as well as an ISDN BRI or DSL line for backup (or a
second T1 routed through a different telco CO if you absolutely need the
bandwidth, but then you'd probably have to be willing to re-locate to a
place in the city where such redundancy would be possible at reasonable
cost).  UUNET should be able to assist in getting full link reduncancy.

> - We canít co-locate our web service to an ISP/Hosting-Provider at the current time.

OK, but why not?  They have the advantage of aggregating many costs
together and thus can provide far more reliable services than you can
ever hope to do at any similar cost on your own.

> - Our current provider, UUNET (weíre using T1 burstable service), has a single POP in
> our location (Ottawa, Ontario Canada).  We donít want 2 links to the same provider (POP), or
> do we?

To decide how "deep" you need redundancy you have to look at where the
risks are.  You also have to have a long and hard look at exactly where
the primary community you serve is located on the Internet.

I'd bet that >50% of the risk lies in your own on-premises facilities.
About 25% of the risk will be the local loop to UUNET's POP, and the
rest of the risk is that UUNET's link to Ottawa will go down.

However the perception you no doubt have is that if your link goes down
you're dead no matter what else might happen.  In that case link-level
redundancy to your provider will suffice to eliminate the obvious
finger-pointing problems.  Then what remains is either entirely your
responsibility (eg. your building burns to the ground, or your disk
fries and you learn your backups are all garbage); or UUNET's
responsibility (eg. someone digs up the fibre they use to connect to
Toronto/Montreal/wherever).

If your customers are also regional UUNET customers then having
redundancy to another ISP isn't likely going to help you any if UUNET
themselves are down for whatever reason.  If your customers are mostly
in the USA then you should probably think harder about why you haven't
moved your servers to a good co-location facility in the USA.  If your
customers are mostly in Europe, then why aren't your servers there too?
If your customers are all over the place then why don't you have
multiple servers located in diverse (Interent geography-wise) locations?
We're talking about the Internet here aren't we?  It shouldn't matter
one iota where your servers are located!

Before you say that you must have 24x7 availability you really need to
think awfully hard about just how much money that level of service is
worth to you, and then you have to get some expert advice to tell you
how much that level of service costs in the real world.  If the numbers
don't match then you really need to carefully analyze the risks and the
costs to mitigate them and then find some balance between what you can
afford to spend and what you can do to reduce the biggest risks equally
across the board.

> - As far as I know there is only one provider in this area that has > 1 POP. [its not
> financially
> prudent for us to drop UUNET and go with them \226 read \205 we have a \223contract\224].

Oops -- sounds like someone didn't do enough up-front planning for this! ;-)

> - We have been allocated a /24 from UUNET.

IP allocation is essentially meaningless in your case.  You are not
going to benefit from any kind of IP routing redundancy unless you can
pull your own fibre down different routes to different locations in the
USA.  Period.  Don't even think about BGP -- it won't help you unless
you're willing to pay mega-bucks for your own long-haul links and unless
you've got incredibly "secure" facilities.  "multi-homing" to two
different providers in the same city isn't really going to be any more
reliable than simply paying one good provider enough incentive to sign a
decent service level agreement with you and let them deal with the
redundancy issues to the rest of the Internet.

(Does your UUNET contract include a service level agreement that
reflects your true requirements?  If not, why not?)

> BTW: Since last Thursday afternoon, we've been having T1 link flaky-ness and have been
> intermittantly
> up and down (mostly down).  We replaced our router [it was delivered to the guy in charge of
> our network
> on Saturday night - sort-a link a pizza delivery - One router please, with a T1 for extra
> topping...:-)] and the
> T1 card and some cabling - the finger point is still fricken going on.....So please don't try
> to tell me that us
> small /24 guys don't need link redundancy and multi-homing....we do...

Multi-homing != link-level redundancy.  Physical multi-homing is one
hell of a lot cheaper than IP multi-homing.  Fix the right problem!
Don't let "teething problems" fool you into thinking you need something
that you don't.

-- 
							Greg A. Woods

+1 416 218-0098      VE3TCP      <[email protected]>      <robohack!woods>
Planix, Inc. <[email protected]>; Secrets of the Weird <[email protected]>