Hi All

My current POC setup is as follows. The idea is to use a group identifier specific to each client. Is this a typical/recommended usage suitable for scaling?
- One display table which is dynamically generated with unique “row=X” for each price, but same “table=pricing” attribute for all prices.
- One MetaPush table with a string group identifier that is unique for each web client, eg. group=”price_<sessionId>”
- On the server side, prices are published to light streamer (via remote adapter) using the unique KEY=”price_<sessionId>”.

An alternative is to use multiple group identifiers specific to each market (for pricing) and not specific to each client. Is the following setup more efficient/recommended?
- Multiple MetaPush tables, one for each price. Group identifier is unique for a price, but not unique to each web client (ie. does not include session id). Eg. group=”price_<marketId>”
- At the server side, prices are published only once to LS (via remote adapter) irrespective of the number of subscribed web clients. Thus presumably, LS server would in turn send a price update to each subscribed (if any) client with a matching key=”price_<marketId>”.

Or, should I be doing something totally different? Perhaps using a GroupDescriptor/GroupLineDescriptor?

Many thanks.