Boston Linux & Unix (BLU) Home | Calendar | Mail Lists | List Archives | Desktop SIG | Hardware Hacking SIG
Wiki | Flickr | PicasaWeb | Video | Maps & Directions | Installfests | Keysignings
Linux Cafe | Meeting Notes | Blog | Linux Links | Bling | About BLU

BLU Discuss list archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Discuss] Off-Topic [IP] BufferBloat: What's Wrong with the Internet?

Rich Braun wrote:
> Bill Horne:
>> At some point, the Internet will need a major overhaul.
> Will it?

I think we have a very long history of incremental tweaks ahead of us...

>> For what common carriers are trying to do...TCP/IP can't be made to fit.

I'd buy that TCP may not be part of the future for real-time
communications - heck, it hardly is now, given that Skype, SIP, RTP are
all UDP - but I think what you intended is that the packet-switched
infrastructure used by IP isn't workable.

On that I disagree, as the desire to provide super cheap communications
is too great. Unless a cheaper alternative comes along for the OSI layer
2 and 3 infrastructure[1] we currently have, we'll see creative
solutions at layer 4 (TCP, UDP, SCTP[2]), or just faster pipes, as Rich
suggests - whatever ends up being cheaper to implement.


Providers are currently racing at breakneck speed towards zero cost
telecommunications. Consider Google Voice, or Republic Wireless (whose
parent,, provides the infrastructure for Google Voice,
Skype, etc.), which obtained its own country code so your friends and
family in other countries can call your Republic Wireless phone for free[3].


As phone service becomes like a disposable commodity, people will be
more tolerant of reliability problems, and likely will have a variety of
alternate channels to choose from if one doesn't work to their liking at
the moment.

Also consider that many of us are increasingly using less and less
real-time communications. Replacing phone calls with text messages and IMs.

>> This fight will be
>> about which mega-corporations carve out virtual slices of Internet
>> bandwidth so that they can avoid paying for their own.

There will likely emerge premium services that give you guaranteed
latency, using things like RSVP[4] or NSIS[5], but I think the vast
majority of users will find the commodity service to be "good enough."


> The gearheads who recognize BufferBloat will ultimately do the
> obvious: crank down the buffering and adjust the retry parameters.
> And flatten out the number of hops from source to destination.


[Thanks to Stephen Ronan for sharing the article. I'd heard about this
buffer bloat issue before, but hadn't read the details.]


Tom Metro
Venture Logic, Newton, MA, USA
"Enterprise solutions through open source."
Professional Profile:

BLU is a member of BostonUserGroups
BLU is a member of BostonUserGroups
We also thank MIT for the use of their facilities.

Valid HTML 4.01! Valid CSS!

Boston Linux & Unix /