# MapReduce Pool Model

## Building MapReduce Model
github : [MapReduce Model](https://github.com/ilkclord/Map_Reduce)
### Schedule
* [x] jobqueue
* [ ] Main Pool
* [ ] Error Handle

### Jobqueue
For the jobqueue , I choose ringbuffer as the data structure .There are some advantages and will be introduced below .
* **Accuracy**
Compared with linked list , the `add` / `remove` option in ring buffer is more easier to maintain . If we can assure the `read` / `write` in a correct position , then there won't arise an Error .
* **Fixed Size**
Since the ring buffer limits the entry of the elements , we can implement by using an array which only cost us const time for finding the specific data .

#### Position
When `queue_r` or `queue_w` were called , we reserve a position for them ,and keeps going on .
#### Flags

To assure the write / read operation isn't interrupt , I use 3 flags to show the states and assure the accuracy .
`readable` :　the node is ready to be read , can't be written
`writing` : the node is hooking right now
`raw` : the original state , is ready to be written

For the termination of the queue , cause we have a while loop to wait the node ready to be written or read , we use flags to break out the loop

`act` : queue is running now
`rc` : let the reading process terminate (stop waiting_
`wc` : let the writing process terminate (stop waiting)
`rest` : queue isn't running now

Note that the sequence of calling `rc` and `wc` matters , if `wc`->`rc` , we expected the data were all read out the queue . If `rc`->`wc` , we expected the data were all written into the queue .

### Main Pool
***upcoming ~***