Thank you for the constructive reply. I appreciate it very much.
-Louis
James Carlson wrote:
> Siwei Liu - Sun Microsystems - Beijing China writes:
>
>> James Carlson wrote:
>>
>>> OK ... though there are likely a few controversial issues buried in
>>> there.
>>>
>>>
>> did you mean potential impact of over-stressing the network or servers?
>> I'd like to hear your concerns and comments here.
>>
>
> That's one of them. The other is that when TCP is working correctly,
> there shouldn't be any need for something like this. TCP's designed
> to fill the pipe.
>
>
>> If so, we could apply a patch against aget to limit the
>> connections/pthread to a safe value. Or any other similiar mechanism to
>> minimize the side effect.
>>
>
> There's not much point; the malicious user can always launch a flurry
> of processes to get around any limit at the application UI level.
>
>
>>> The unfortunate thing here is the apparent long-standing disagreement
>>> between the aget and wget folks on multiple streams, leading to
>>> duplicate tools. *sigh*
>>>
>>>
>> Better late than never. :-)
>> PS, any chance for peer to peer software to integrate into Solaris at
>> present? I ask this just for my own interest. I think Transmission is
>> the one for bittorrent network for the time being. How's the strategy of
>> Solaris over the multiple stream software for now?
>>
>
> This likely isn't the right list to discuss that. I suggest
> networking-discuss at opensolaris.org instead, or perhaps some internal
> list instead.
>
> If you're asking about the architectural review committee's position
> on that software, I don't think we have one. As an ARC member, I
> can't see an obvious problem with including it, but I'd have to see a
> project in front of me for review to comment further.
>
> If you're asking about which one of the implementations is "best" or
> how someone would choose among them, then that sounds like something
> for the Networking Community to discuss. The ARC typically doesn't
> get involved in those discussions until there's a decision (that is, a
> project) to be reviewed. (Though I suppose that a project -could-
> list alternatives for an inception review ...)
>
> If you're asking about resources (personnel) or strategy for Sun's
> products, then that's something best asked internally, and it has
> nothing to do with the ARC, and perhaps not much to do with
> OpenSolaris. At a guess, it's something for your management and/or
> for the Solaris PAC.
>
>
--
Best Regards
______
/_____/\ Si-Wei Liu
/____ \\ \ Sun Microsystems China (ERI)
/_____\ \\ / Email: Si-Wei.Liu at Sun.COM
/_____/ \/ / / Direct: (86-10)6267-3670
/_____/ / \//\ SWAN: 51670
\_____\//\ / /
\_____/ / /\ /
\_____/ \\ \
\_____\ \\ 7/F, Tower A, Tsinghua Science Park
\_____\/ Beijing 100084, P.R.China
