Re: Defeating Firewalls: Sneaking Into Office Computers From Home

roberson@ibd.nrc-cnrc.gc.ca (Walter Roberson) wrote:
>People around here don't buy T1's for -speed-. Reasons why they
>do buy T1 include:

This is an interesting list, and technically is correct in most
aspects (the comment about "synchronous ... for multimedia"
isn't right).

However, there are some *very* serious "gotcha's" in this too.
Most services provided over what might be called a "T1", may or
may not be technically what is being reference as a T1!

For example, a T1 that is provision through a provider's Frame
Relay cloud, might not provide the stated benefits. That isn't
technically just a T1, but is almost always called a T1 today.

>- T1 provides symmetric bandwidth (SDSL is not very common here)

Significant, and correct for anything likely to be called a T1.

>- T1 is point to point and thus does not have the security issues
>that one has when one is connected to the Internet

That is probably not quite true, though the issues are less
serious than with DSL or Cable provisioned services. A Frame
Relay Cloud still has the same potential insider security
issues...

>- T1 are available -most- places that phone land-lines reach --
>the required infrastructure having existed for long enough for
>very good saturation.

True.

>- T1 do not degrade with congestion from other users on your block
>(a significant problem with cable!)

Unless of course it is part of a Frame Relay Cloud. (I once saw
a carrier install a Frame Relay Point-of-Presence, and then sell
6 different customers guaranteed bandwidth circuits equal to the
backbone! ;-)

>- T1 have fixed latency and bandwidth, not variable as cable has

Probably true, but it can vary significantly if any part of the
T1 is on Frame Relay or Cell Relay (ATM) backbone services.

>- T1 is synchronous and thus suitable for multimedia applications
>that degrade when frames are received out of order

Being synchronous or not doesn't affect the order that frames
are received in. Regardless, virtually all T1's ride T3's, and
thus become isochronous and suffer from higher clock jitter than
a genuine point to point T1 over a wire facility would have.

>- T1 has no bandwidth caps, content filters, forced http proxying,

Two out of three. It certainly has a bandwidth cap. Moreover
it it is provision on Frame Relay or Cell Relay there will be
just as many wierd bandwidth specs as one can imagine (burst,
guaranteed, etc.), and might well be affected by how much other
customers are using.

-- 
Floyd L. Davidson            <http://www.apaflo.com/floyd_davidson>
Ukpeagvik (Barrow, Alaska)                         floyd@apaflo.com