[NikitaP]

Anand,

1) CloudSearch does support utf-8, actually there are stores running CloudSearch with slovenian, greek, french and other languages successfully.

2) At the moment there are plenty of server resources dedicated to support CloudSearch growth. We use server and application monitoring to control the performance. Whenever we spot performance degradation we add more server resources. Speed is one of the main features of CloudSearch and we are not going to give up on this.

3) As long as urls of the images are the same it doesn't matter how they were assigned.

4) It is a good question. When we designed CloudSearch we considered two approaches: to periodically re-index all the data or do it incrementally. We took the first approach because of simpler connector implementation. Incremental approach requires implementation of some (relatively) sophisticated logic on the connector's side to track changes to searchable entities, and it will put some additional load on the client's server. There would be some performance gain on our side but I can't say that incremental approach would be more effective for connector implementation (not to mention reliability). The load CloudSearch puts on X-Cart is very low because of simple SQL queries and little processing. There is always a dilemma: do something on one side of the system or on the other. The main approach we take is to do more on our side and less on X-Cart side. As long as I see this approach justifies itself.


5) CloudSearch considers product name, description, SKU, manufacturer, product options names, extra fields and category name for each product. Keywords are not considered. We are trying to use only the data that is shown to customers.

There are also static pages, manufacturers and categories - they can be searched too.

6) Not really, but this can surely be done by customizing CloudSearch connector code.


I highly encourage you to try free evaluation to see it all in action.