Blitline lets you run massively parallel automated image processing jobs in the cloud.
First you must log into Blitline and get a FREE developer account. When you register, you will get an Application ID which is your secret token you will use for all you processing jobs.
Clients submit JSON to Blitline defining image source and the sequence of operations that one wishes to perform on the given source image. You can pipeline many different operations and output to many different locations and filetypes.
There is a single primary endpoint which you can submit your JSON to. This is a POST to:
OR for processing in EU (Frankfurt)
The scope of the JSON you submit is larger than we will dive into here, but it is based on the notion of a “job”. This “job” is an atomic entity which encompasses all the operations you wish to perform on a single image. You can have multiple “jobs” in a single submission, but you must have at least 1 job to perform any actions.
Scalable applications require layers that absorb load and distribute that load across workers that can be scaled up and down as that load ebbs and wanes.
When jobs are submitted to Blitline, they are put into distributed queues that hold the data until backend servers can pick them up and process them. This processing (on average) takes about 30ms, but that can vary based on load.
When jobs are picked up by Blitline workers, they are executed, images produced and output, and postbacks(webhooks) called and clients notified.
For this process to function well, it is optimum for clients to define a postback_url which is a URL of the client’s that Blitline can call to notify them that the job has completed. While it IS possible to poll for completion, it is not recommended for production.
You can see a prezi presentation describing Blitline workflow here
CAN I UPLOAD IMAGES TO BLITLINE?
Not many people would argue that S3 or Azure are probably the largest, fastest, best outfitted platforms for uploading images to. It is not reasonable to think that Blitline could reach a level of upload performance that these platforms have, so we have decided there is little need for us to try to compete in this space.
Updated over 1 year ago