Amazon SQS - JMS vs custom implementation -

currently reading sqs queue following code

public class sqsreceiver implements runnable {      @autowired     protected amazonsqs sqs;      protected receivemessagerequest receivemessagerequest;      public void run() {         list<message> messages = sqs.receivemessage(receivemessagerequest)                 .getmessages();         (message msg : messages) {             logger.info("processing message: {}", msg.getbody());             try {                 processmessage(msg);             } catch (exception e) {                 logger.error("error processing message", e);             } {                 sqs.deletemessage(new deletemessagerequest().withqueueurl(                         receivemessagerequest.getqueueurl()).withreceipthandle(                         msg.getreceipthandle()));             }         }     } } 

and having scheduler read every second:

threadpooltaskscheduler scheduler = new threadpooltaskscheduler(); scheduler.setpoolsize(1); scheduler.initialize(); scheduler.schedule(sqsreceiver, new crontrigger("* * * * * *")); 

but recently, learned amazon released using jms amazon sqs. wondering besides having code independent underlying implementation, benefits there in terms of performance , costs in switching implementation use jms?