The response time of any service in your application depends on what the service is doing. If you create a service that searches tables in pega but uses SQL that creates table scans or returns millions of rows then your service wont respond in less than a second.
Services have build in alerts so that if processing takes over x amount of time for each major part of the processing you can tell what is taking the time:
PEGA0011 (1 second) - Total request time exceeds limit (Service Total Time)
PEGA0015 (500 ms) - Data parsing time exceeds limit (Service Parse Time)
Based on the above and too answer your question we expect that you will create a service that runs in less than one second. If the service doesn't run in less than a second you have alert data that shows you what took the time. When you create your service you should test it with expected load and and tune as needed.
If the service you create requires authentication to run then each time a client calls the service the service requestor will have to go through authentication layers. This can add more to the processing time. If you don't use authentication you can use pools of requestors to process the requests.