Let us all bend over, apply the Vaseline...

[Jon Lewis]

On Sat, 3 May 1997, Michael Dillon wrote:

> > I personally think that if the telco's
> > and other giants ever get serious about providing internet service with
> > clues, many (but certainly not all) smaller ISP's will be unable to
> > compete.
> 
> That's a lot of ifs. Many people feel that it is simply not possible for a
> large company in any industry to maintain a high enough clue level to
> compete for very long against smaller entrepreneurial companies. I tend to
> agree with this. However I also agree that smaller ISPs that do not have a
> clue will be unable to compete with anyone.

The thing is, if they have enough money, clues are optional.  Bellsouth
just decided to increase our phone bill by 20%, by adding new fees for a
service used primarily by ISPs.  This is the sort of big business
squeezing the little guy that the new telco/ISP's are using to try to
drive smaller ISP's out of business while increasing their own market
share. 
 
> > > Those that do stay in business will require the reliability of a
> > multihomedmed internet connection.
> 
> I used to think that most ISPs would become multihomed but as I've watched
> things evolve over the last year and a half, I'm no longer so certain that
> this will happen or that it will be necessary. Multihoming is only one
> part of the redundancy equation in which an ISP strives to make sure their
> connectivity into the global Internet is redundant enough that no single
> point of failure can cut them off completely. The idea is that having
> multiple backbone providers protects you from screwups by any one of them. 

FDT has a single T1 link to UUNet.  That T1 has been the single largest
source of downtime type problems we've had to face.  During one 3 month
period, we had several dozen hours of downtime with the longest outage
lasting well over 12 hours.  I'd have to go back and check logs, but I
think it was 17h.  We got no explanation for most of these outages.  The
scenario was typically:
T1 goes out.  We call UUNet to report loss of service.  They look into it
and agree.  Sometimes over an hour would go by before they call Worldcomm.
Time goes by, and eventually Worldcomm says their part of the loop is
clean, so UUNet reports the problem to BellSouth (local loop provider).
BellSouth generally takes a while, but eventually reports their part of
the loop is good.  While everyone's playing phone tag pointing fingers,
the T1 magically comes back up.  Nobody claims to have a clue what caused
or fixed the problem.

The few outages we did get explanations for were chronic problems in a
cascade switch at UUNet's JAX POP.  The Switch stops working, we lose
service, and UUNet eventually replaces a card in it and it works again.
Then the whole unit starts acting up and UUNet announces a planned outage,
skips it, but then does it the next night without warning.

During the particularly bad 3 month period, I talked to UUNet about doing
a backup ISDN link for times that our T1 was out for extended periods, and
the engineers would agree, but say to talk to sales...and sales would
always say "we can't do that".

Well...you can't rely on a single provider like this forever and expect to
stay in business.  Each time we've had an extended outage, we've lost a
few customers our reputation has suffered.  If renumbering weren't such a
bitch, we'd probably have left UUNet late last year.  There will come a
time (soon, I hope) when we can afford the hardware necessary to
effectively utilize a second link to the net.  The more we have to pay
ARIN, the longer it will take us to get there. 

> So I think that in the grand scheme of things, multihomeing is less
> important for the smaller ISPs than redundancy. And there are many ways in

Nobody plans to stay small forever (at least in the ISP biz).  FDT's been
doing this for two years, and I'd like to think we're not still "small".
Certainly there will always be smaller ones, for whom multiple connections
to top level providers just isn't feasible...and some of them will/do get
connections from us, thus increasing the need for our connection to the
net to be more reliable.

If anyone's still reading at this point, I'd just like to say again and in
public, I'm not oposed to ARIN.  I'm not totally opposed to the fees,
though I'm not happy about them.  And I think any organization that pays
for an IP block allocation should be given a free membership in ARIN for
as long as they hold that block.  i.e. no $1000/year membership fee for
organizations that are actively using ARIN allocated space and thus have a
greater interest in how ARIN is run.

------------------------------------------------------------------
 Jon Lewis <jlewis at fdt.net>  |  Unsolicited commercial e-mail will
 Network Administrator       |  be proof-read for $199/hr.
________Finger jlewis at inorganic5.fdt.net for PGP public key_______