Hi Lily is an indexing solution for HBase. This indexing is purely happening at the client side. If you see Lily will sit in between the client app and HBase. The app need to insert/delete data via Lily only. Lily will write the user data into HBase table. Also there is another index table which it created for storing index details. Lily write this index data too. While deletion of it will handle both the table. Scans also to go via Lily only. It will fetch the index data 1st and accordingly will fetch required rows only from the actual ctable. Hope this makes it clear for you.. :)
-Anoop- ________________________________________ From: Jason Huang [[email protected]] Sent: Friday, September 28, 2012 2:28 AM To: [email protected] Subject: HBase and Lily? Hello, I am exploring HBase & Lily and I have a few starter questions hoping to get some help from users in this group who had tried that before: (1) Do I need to post all the HBase table contents to Lily (treat Lily as another DataStore) in order to enable the index and search functionality? If so, that's going to be another big storage issue (and duplicate storage?) (2) Should I always only allow one way update from Clients -> HBase -> Lily? I want to use HBase as the data store and use Lily only as a plug-in tool to help search. I want HBase to only accept updates from Clients (not from Lily). Is there any update from Lily to HBase required (in order to enable the search and index functionality)? (3) Since Lily is not an Apache project - do you know if it's under Apache 2.0 license? We may need to extend it with our own APIs. Do we "have to" give our APIs back to them? We love sharing but some of our APIs may be under different agreements and can't be shared. thanks! Jason
