IMs and group IMs (was Re: [sldev] Frequent bugs with difficult repros)

[Tateru Nino]

Henri Beauchamp wrote:
> On Fri, 29 Feb 2008 17:11:05 -0700, Lawson English wrote:
>
>   
>> Tateru Nino wrote:
>>     
>>> Kelly Linden wrote:
>>>       
>>>> Tateru Nino wrote:
>>>>         
>>>>> Soft wrote:
>>>>>           
>>>>>> On Fri, Feb 29, 2008 at 2:51 AM, Callum Lerwick <seg at haxxed.com> 
>>>>>> wrote:
>>>>>>  
>>>>>>             
>>>>>>>  On Fri, 2008-02-29 at 13:24 +1100, Tateru Nino wrote:
>>>>>>>  > Double chats in IM groups? That's caused by packet loss between 
>>>>>>> the
>>>>>>>  > viewer and the sim, honey. Same as double-chat in local chat, 
>>>>>>> and in
>>>>>>>  > person-to-person IMs. It also causes out-of-order chat.
>>>>>>>
>>>>>>>  So when is our chat going to be transported by this new-fangled TCP
>>>>>>>  thing? I hear it deals with packet loss and retains ordering...
>>>>>>>     
>>>>>>>               
>>>>>> I'm pretty sure TCP is just a fad.
>>>>>>
>>>>>> Actually, I thought this had already gone TCP. I'll ask around, but if
>>>>>> any other Lindens want to chime in...
>>>>>>   
>>>>>>             
>>>>> Depends. Even if it is tcp, application level sequencing and 
>>>>> controls over messaging flow, acknowledgement and order can still 
>>>>> duplicate your data - depending on whether you're second guessing 
>>>>> the transport with your own scheme, and how you've got it implemented.
>>>>>
>>>>>           
>>>> Those would be what we would call "bugs", I hope.  :D  I'm not the 
>>>> most up to date (my head is far too deep in havok4 for the past 6 
>>>> months), but I don't think chat or IM have moved to TCP, I believe 
>>>> they still use our home brewed "reliable" UDP messaging transport layer.
>>>>
>>>>         
>>> I believe that is also the case. Reliable is such an interesting word. 
>>> Very contextual. :)
>>>
>>> That said, the algorithms that govern the UDP transport actually do a 
>>> surprisingly good job under very trying and harsh circumstances - 
>>> however circumstances and workloads just aren't what they once were. 
>>> Something I think we all recognize.
>>>
>>>       
>> As far as I know, the main issues with IM in Second LIfe have nothing to 
>> do with UDP and would be just as bad if they were implemented using TCP, 
>> as long as the same underlying design remains unchanged.
>>     
>
> The IMs and group IMs issue is indeed independant of the protocol. IMs
> used to work just fine in UDP before LL made one of its most catastrophic
> mistakes by changing the architecture for IM communications in v1.14.1
> ( http://blog.secondlife.com/2007/04/02/preview-of-second-life-11411-now-up-on-the-beta-test-grid/ )
> to try and lighten the load on the servers.
>
> Despite a HUGE amount of bug reports and complaints, LL stubbornely kept
> the changes they made and we are still now, 11 months after the problems
> were identified, suffering the same out of order IMs, spuriously reopening
> group sessions, failure to open group sessions and send group notices,
> failures to set invisble or visible status in friends list (you are marked
> either visible or not, but your friends see the opposite), failures to load
> friends or groups lists after a cache emptying, etc, etc, etc...
>
> LL, please, STOP this madness and return to the OLD, reliable architecture,
> simply using better pipes and more powerful servers to keep the IMs and
> groups CENTRALIZED !
>   
Group IMs indeed have their problems - but I consider those issues to be 
quite distinct from the others. Not that they are any less important - 
they're just a completely different class of problem.

-- 
Tateru Nino
http://www.massively.com/