Thanks! I got the "s3" blobstore working (using the official AWS SDK) and I 
am finishing up the "dynamodb" key/value store.

After that I should be ready to go and I may need some ideas on how to 
properly implement authentication for this (I could easily do a "public" 
blobserver, but it wouldn't be very useful)

-- Raffaele



On Saturday, December 10, 2016 at 10:29:29 PM UTC-8, [email protected] 
wrote:
>
> Hey. I used to be an Amazon Engineer on the AWS Lambda team, I can help 
> provide guidance if you want some. :)
>
> On Wednesday, 23 November 2016 11:08:09 UTC-8, Raffaele Sena wrote:
>>
>> I would like to use Camlistore for my pictures, that I rarely browse 
>> until I need them (I am currently hosting most of them on a VPS and I don't 
>> remember when was last time I looked at them) and I thought I'll store them 
>> in S3 since I am already using AWS for other purposes.
>>
>> I know I could run a local server configured to use S3 but I do sometime 
>> access the pictures from random browsers (and maybe for that use case I 
>> could just run camlistored on a micro instance and that would be good 
>> enough).
>>
>> But I have been toying with the idea of running the blob server out of 
>> AWS Lambda and I am trying to figure out if this makes any sense or not (if 
>> it does maybe the next step would be running other parts of camlistore on 
>> Lambda).
>>
>> Basically I was going to start from the "appengine" code (
>> https://github.com/camlistore/camlistore/tree/master/server/appengine) 
>> and tweak it to run on AWS.
>>
>> Somebody recently created a package to run Go application "natively" on 
>> AWS lambda (no JS or python proxy) so there would be no overhead there.
>> And Lambda services once started do run for a while (not sure of how long 
>> they run, and if there is a way to get a signal when they are terminated, 
>> but I am doing some experiments to figure that out).
>>
>> I am still debating if there would be more overhead running a local 
>> server that push data directly to S3 vs. pushing blob to the hosted blob 
>> server but again one of the purpose of this exercise would be to have the 
>> full environment running in AWS "on demand".
>>
>> So, is this a crazy idea that I should abandon right now or does it have 
>> some value ? Is there anything major that I am missing (I didn't look at 
>> the code that much but I am assuming the blob server should keep a "global 
>> state" that requires for the server to be running all the time, and the 
>> overhead of starting the process is relatively small.
>>
>> Thanks!
>>
>> -- Raffaele
>>  
>>
>>
>>
>>
>>

-- 
You received this message because you are subscribed to the Google Groups 
"Camlistore" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to