| 64 |
|
* <dt>Queueing</dt> |
| 65 |
|
* |
| 66 |
|
* <dd>You are free to specify the queuing mechanism used to handle |
| 67 |
< |
* submitted tasks. The newCachedThreadPool factory method uses |
| 68 |
< |
* queueless synchronous channels to to hand off work to threads. |
| 69 |
< |
* This is a safe, conservative policy that avoids lockups when |
| 70 |
< |
* handling sets of requests that might have internal dependencies. |
| 71 |
< |
* The newFixedThreadPool factory method uses a LinkedBlockingQueue, |
| 72 |
< |
* which will cause new tasks to be queued in cases where all |
| 73 |
< |
* MaximumPoolSize threads are busy. Queues are sometimes appropriate |
| 74 |
< |
* when each task is completely independent of others, so tasks cannot |
| 75 |
< |
* affect each others execution. For example, in an http server. When |
| 76 |
< |
* given a choice, this pool always prefers adding a new thread rather |
| 77 |
< |
* than queueing if there are currently fewer than the current |
| 78 |
< |
* getCorePoolSize threads running, but otherwise always prefers |
| 79 |
< |
* queuing a request rather than adding a new thread. |
| 67 |
> |
* submitted tasks. A good default is to use queueless synchronous |
| 68 |
> |
* channels to to hand off work to threads. This is a safe, |
| 69 |
> |
* conservative policy that avoids lockups when handling sets of |
| 70 |
> |
* requests that might have internal dependencies. Using an unbounded |
| 71 |
> |
* queue (for example a LinkedBlockingQueue) which will cause new |
| 72 |
> |
* tasks to be queued in cases where all corePoolSize threads are |
| 73 |
> |
* busy, so no more that corePoolSize threads will be craated. This |
| 74 |
> |
* may be appropriate when each task is completely independent of |
| 75 |
> |
* others, so tasks cannot affect each others execution. For example, |
| 76 |
> |
* in an http server. When given a choice, this pool always prefers |
| 77 |
> |
* adding a new thread rather than queueing if there are currently |
| 78 |
> |
* fewer than the current getCorePoolSize threads running, but |
| 79 |
> |
* otherwise always prefers queuing a request rather than adding a new |
| 80 |
> |
* thread. |
| 81 |
|
* |
| 82 |
|
* <p>While queuing can be useful in smoothing out transient bursts of |
| 83 |
|
* requests, especially in socket-based services, it is not very well |
