[vlc-devel] Streaming wizard issues

[jpd at videolan.org]

On Mon, Nov 30, 2009 at 05:44:18PM +0100, Marian ??urkovi?? wrote:
> It doesn't limit the damage, it creates one, even with biggest routers.

Who has a linkup to the biggest routers and _doesn't_ know to set this
straight? The point is to stop essentially unconfigured stuff right at
the edge. As long as people who don't know what they're doing remain at
the edge, TTY=1 does that admirably.

People who have such linkups presumably have read the RFCs, have
communicated with or are their network administrators, and know what
scoped addresses are in use, and thus know what TTL to override the
default with.

Remember that nobody proposed to use this in production; rather to make
sure unconfigured setups don't accidentally cause problems in production.

Your complaint of overloaded core routers would happen if we'd set the
TTL to some arbitrary number above one but below the average diameter of
the internet, and people all over the internet would start to try and
stream multicast using our defaults. With TTL=1, that does not happen,
and if it does the culprit is on the same link rather than several ASen
away. If someone accidentally DoSes his own gateway he has a clear
incentive to stop. If he does it to someone else's backbone, he may not
even notice.


> It violates the current RFCs. 

Er, no, it doesn't. "SHOULD" is not quite the same as "MUST". It would
violate the SDP RFC only if it said TTL "MUST NOT" be used this way, but
it explicitly does not. You know what RFC to look up for the meaning of
the RFC-legalese.


> And it confuses users who in turn generate unnecessary support queries.
> That could hardly be called sane, IMHO.

``Unnecessairy'' is debatable. Better to cause a first level support
ticket and fifteen minutes of impromptu user education than an escalated
ticket followed by a few hours tracing an obscure multicast problem,
in turn followed by a shouting match between an irked network tech and
an unhappy user, I say. With an educated user who knows to configure
for the scoping in use at that site neither should happen. Debating the
sanity of, as is currently the mantra in computing, refusing to educate
users *before* letting them loose on the network is a whole 'nother
kettle of fish, and somewhat outside of the scope of this discussion.