[Cryptography] Apple's iCloud+ "VPN"

Natanael natanael.l at gmail.com
Wed Jun 30 03:47:03 EDT 2021


On Wed, Jun 30, 2021 at 7:41 AM Bill Woodcock <woody at pch.net> wrote:

>
> And that puts us right back to what I said in the last email.  There’s no
> magic here.  Adding two new segments to a path cannot decrease the length
> of the path.  There’s no such thing as a negative-length segment.
>

Private links is not the only counterargument. Remember when Comcast
throttled Netflix indirectly by routing all traffic to the same
underdimensioned interconnect point? At that point you could in fact get
higher connection speeds by using a VPN, since that would direct you
through an interconnect with higher capacity. There are also other examples
of traffic simply being routed inefficiently, not taking the shortest
available path even if that path would otherwise be able to handle the
traffic.

Yes, this is only relevant for throttled or poorly routed traffic, but this
is still a thing which happens in the real world even if relatively rare.
But when it happens, you can get better performance by using an alternative
route. Yes, the average case probably means higher latency. Still
occasionally possible.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://www.metzdowd.com/pipermail/cryptography/attachments/20210630/4ea0d419/attachment.htm>


More information about the cryptography mailing list