In this chapter, we have discussed exactly what we mean by low latency. Often, this in itself is the source of much confusion. You will now understand the effects that the BizTalk architecture has on latency, since its primary use case out of the box is for high throughput. Understanding the architecture as described in this chapter will help you to exploit it to the fullest and to squeeze the most out of it for your scenario. If BizTalk cannot be tuned to meet your latency requirements, the adoption of an inline send port approach is most likely your best bet to meet your targets. The code samples in this chapter will help you to understand how this can be achieved relatively easily and also help you to design your solution so that you don't lose many of the benefits provided by using send ports.
Finally, this chapter has helped you to understand the various options for how you can measure latencies. Performance counters can be used for a coarse-grained view. Alternatively, BAM can be exploited to provide a message-level view of the latency characteristics of your system.