[voeventnet] Re: URLs?

[Joshua Bloom]

Matt et al.

   This sounds great. The main thing idea about decentralization is  
not only that the aggregators have multiple and redundant feed  
sources, but that the subscribers have redundant pushers. We'd be  
happy to try to replicate this aggregator that you are building at  
Berkeley so that subscribers can have at least two sources of VOEvents.

Josh

On Dec 23, 2005, at 11:58 AM, Matthew Graham wrote:

> Joshua Bloom wrote:
>
>>
>> On Dec 8, 2005, at 3:25 PM, Alasdair Allan wrote:
>>
>>>
>>> I think we've moved to a much more decentralised network, and in  
>>> my  opinion this is a good thing! Rob was taking notes during the  
>>> mini- interop and I think we should wait for him to push these to  
>>> the  group before moving forward with this discussion to bring  
>>> everyone  up to speed with what we discussed there (we got  
>>> through a lot of  stuff).
>>>
>>
>>    I am pleased to know that we're moving to a "much more   
>> decentralised network." Getting feeds from multiple sources, even  
>> if  there is substantial overlap, is the only way to insure proper  
>> follow- up. For those of you not working on GRBs on a day to day  
>> basis, you  might not be aware of a fault that occurred on Dec 21  
>> UT with the GCN  socket notifications. Though it was quickly  
>> repaired, the community  missed out on learning of the RT  
>> localization of one of the brightest  bursts detected by Swift  
>> that was of a very special type ("short- hard"). I triggered the  
>> PAIRITEL telescope by hand once reading a  human generated  
>> circular but we would have observed the source  seconds after the  
>> burst, not 1 hour after had there been multiple  feeds to the  
>> telescope.
>>
>>   The GCN has amazing up time (~99%) and it is simply bad luck  
>> that  something went wrong when one of the monster bursts  
>> occurred, but I  think we need to view this instance as a strong  
>> impetus to remove  single failure points from VOEventNet.  To  
>> start the ball rolling on  this point, I had asked Nat Butler in  
>> November to discuss with the  HETE team the possibility of them  
>> autogenerating VOEvents and  distributing them directly from the  
>> **ground stations** (in addition  to sending info to Goddard).  
>> I've begun proselytizing to other major  event providers (both  
>> current and future) to this effect. Once a  network of aggregators  
>> and relay boxes exists, this will insure that  those that want to  
>> hear of VOevents will, with as much reliability as  the internet  
>> can provide. Again, looking into the crystal ball, I  suspect that  
>> the VOEventNet topology will/should look like the  architecture of  
>> the internet itself (*).
>>
>>   Apologizes for my long radio silence. Nat Butler speaks great   
>> things about the Tuscon meeting!
>>
>> josh
>>
>> (*) http://images.google.com/images?q=internet%20topology
>
> Hi Josh,
>
> We're prototyping a repository for VOEvents here and are trying to  
> connect up to as many feeds as there are to remove this reliance on  
> potential single-point failures. There is a form-based search  
> interface to allow you to retrieve specific packets, see which  
> packets cite a particular packet, do a spatial search for packets,  
> i.e. retrieve all packets within a particular RA and Dec range and  
> do a conceptual search for packets based on the VOConcepts list.  
> You will also be able to subscribe to the repository and only  
> receive packets which match some user-defined filters. I am hoping  
> that we might have a demo/prototype up in time for AAS!
>
>    Cheers,
>
>    Matthew