Slide 16
Slide 16 text
THE BAD
•no message guarantees and clients are responsible for re-queueing
•bottleneck and SPOF transport issue (reconnects, distribution, fault tolerance, etc.)
•inefficiency - we copy streams multiple times to multiple systems
•complicated service setup repeated for each stream
•hard-coded knowledge of queue addresses in queuereaders
•lack of internal performance metrics (clients, rate, etc.)