process_transactions actually implements the 'pull'-way to synchronize transactions between Ripple network and the database. Thus we have to run this command periodically (e.g. by cron). It's inefficient, resource-intensive, slow for response. The end user should never wait even for a minute or two if a transaction has been already completed. Definitely there should be a way to update data on the fly, using some kind of a 'push'. I assume the Ripple client does exactly the same. Maybe also take a look at https://github.com/ripple/gatewayd
process_transactions actually implements the 'pull'-way to synchronize transactions between Ripple network and the database. Thus we have to run this command periodically (e.g. by cron). It's inefficient, resource-intensive, slow for response. The end user should never wait even for a minute or two if a transaction has been already completed. Definitely there should be a way to update data on the fly, using some kind of a 'push'. I assume the Ripple client does exactly the same. Maybe also take a look at https://github.com/ripple/gatewayd