Hi Shreyas,

If you want to take benefit of the highly efficient delivery mechanisms of Lightstreamer (such as delta delivery and message conflation), you should map any data model you have to a table-based data model. So, in this case, it is up to you to find the right mapping.

On the other hand, if you don't need optimized delivery, you can use opaque data types. In other words, you can serialize your complex data object into a string (through JSON, XML, or any other technique) and push it as a single field. Your client-side code will do the opposite operation, deserializing the string and getting a full-blown object again.

Again, consider that by using opaque data types, you actually hide the data model to Lightstreamer, which cannot parse the messages to extract the deltas. So, each message will be delivered "as is" to the client (thus consuming more bandwidth, also due to the overhead of the serialization syntax).

Regards,
Alessandro