1·When mutable state goes away, almost all concurrency problems go away with it.
可变状态消失时，差不多所有的并发性问题也随之消失。
2·Shares some or all of the resources used by the system (i.e., concurrency).
分享一些或所有的系统使用的资源(例如：并发性)。
3·Even in the absence of system failures, there is another risk worth discussing in the above code — concurrency.
甚至在系统没有发生故障时，上面讨论的代码还有另一种风险——并发性。
4·These registry variables can improve concurrency, but can also impact the behavior of the application.
这些变量可以改善并发性，但是也会影响到应用程序的行为。
5·One symptom of insufficient concurrency and insufficient connection pool size is an idle CPU.
并发性不足和连接池大小不足的一个症状是CPU闲置。
1·Increasing the concurrency beyond this value will increase the transaction response time as threads compete for CPU time, and not actually yield any throughput improvement.
当并发数增加到超过该值时，由于线程争用CPU时间，所以事务响应时间将增加，因而实际上并未带来任何吞吐量方面的改进。
2·If our transaction is CPU-bound, then the productive concurrency is likely to be at most two or three times the number of available processors.
如果我们的事务是CPU限制的，则生产并发数可能最多比可用处理器数多两三倍。
3·The Maximum Messages value is set on the listener port, along with the Maximum Sessions concurrency setting.
在侦听器端口设置最大消息值及最大会话并发数。
4·Unfortunately, it is difficult to estimate the amount of concurrency supported by a single dispatcher thread on this kind of system.
遗憾的是，难以评估这种系统中单个调度程序线程支持的并发数。
5·Page Load Time = (Number of Requests / Concurrency) * Latency + Page Total Size / Bandwidth
页面加载时间 = （请求数 / 并发数）* 延迟时间 + 页面总大小 / 带宽