一：问题背景
很多计算任务涉及到海量数据的处理，想要在可以接受的时间内完成计算任务，就必须将这些任务分布到成百上千的机器上运行。
如何分发数据，任务调度，处理容错，这些问题需要大量的代码来处理。
因此实现一个分布式的任务需要处理任务本身的代码+实现分布式的大量额外代码；
为了解决以上问题，MapReduce应运而生。
MapReduce是一个编程模型，隐藏了关于并行计算、容错、数据分布、负载均衡这些细节。
即：用户只用表述想要执行的简单操作，MapReduce可以负责实现自动的并行化和分布式计算任务；
二：编程模型
MapReduce的用户将任务划分为两个计算操作Map() 和Reduce() 。
-
Map()接受输入文件，输出一个 key/value 键值对的集合； - MapReduce模型负责将
Map()函数产生的键值对的集合中，相同的 key 值的value值集合到一起，传递给Reduce()函数。 -
Reduce()接受一个 key 值和相应的 value 集合，合并这些value值，输出一个 key/value 键值对;
统计单词出现次数的示例：
map(String key, String value):
// key: document name
// value: document contents
for each word w in value:
EmitIntermediate(w, "1");
reduce(String key, Iterator values):
// key: a word
// values: a list of counts
int result = 0;
for each v in values:
result += ParseInt(v);
Emit(AsString(result));
三：实现
3.1 执行过程
The MapReduce library in the user program firstsplits the input files into M pieces of typically 16megabytes to 64 megabytes (MB) per piece (con-trollable by the user via an optional parameter). It then starts up many copies of the program on a cluster of machines.
-
One of the copies of the program is special – the master. The rest are workers that are assigned work by the master. There are M map tasks and R reduce tasks to assign. The master picks idle workers and assigns each one a map task or a reduce task.
Task：M+N > Worker
-
A worker who is assigned a map task reads the contents of the corresponding input split. It parses key/value pairs out of the input data and passes each pair to the user-defined Map function. The intermediate key/value pairs produced by the Map function are buffered in memory.
Map阶段：读取文件内容，调用map()函数，写入中间文件；
-
Periodically, the buffered pairs are written to localdisk, partitioned into R regions by the partitioning function. The locations of these buffered pairs on the local disk are passed back to the master, who is responsible for forwarding these locations to the reduce workers.
Map任务成功，返回中间文件的位置信息；
-
When a reduce worker is notified by the master about these locations, it uses remote procedure calls to read the buffered data from the local disks of the map workers. When a reduce worker has read all in-termediate data, it sorts it by the intermediate keys so that all occurrences of the same key are grouped together. The sorting is needed because typically many different keys map to the same reduce task. If the amount of intermediate data is too large to fit inmemory, an external sort is used
Reduce阶段：获取key region的所有中间文件内容，排序生成key-values集合，调用reduce()函数，写入输出文件；
The reduce worker iterates over the sorted intermediate data and for each unique intermediate key en-countered, it passes the key and the corresponding set of intermediate values to the user’s Reduce function. The output of the Reduce function is appended to a final output file for this reduce partition.
When all map tasks and reduce tasks have been completed, the master wakes up the user program. At this point, the MapReduce call in the user pro- gram returns back to the user code.
3.2 Master数据结构
Master存储每一个Map任务和Reduce任务的状态：空闲、工作、完成；以及非空闲任务的worker的机器标示；
Master存储中间文件的位置信息，因此Map任务完成时，对应的中间文件位置信息也会更新，最终传递给Reduce任务；
3.3 Fault Tolerance
-
Worker Failer
master周期性ping worker，超时标记为fail。
这个worker正在运行的map任务或reduce任务将被重置为空闲状态，等待调度；
这个worker已经完成的所有map任务也将重置为空闲状态，等待调度；
其他worker正在运行的reduce任务也将重置为空闲状态，等待调度；已经完成的Map任务文件存储在本地磁盘，节点故障后无法访问，需要重新执行；
已经完成的Reduce任务文件在全局文件系统GFS，节点故障也没关系，不用重新执行；
-
Master Failer
一个简单的解决办法是让master周期性的将上面描述的数据结构写入磁盘，即检查点(checkpoint)。
如果这个master任务失效了，可以从最后一个检查点开始启动另一个master进程。
然而，由于只有一个master进程，master失效后再恢复是比较麻烦的，因此我们现在的实现是如果master失效，就中止MapReduce运算。客户可以检查到这个状态，并且可以根据需要重新执行MapReduce操作。
3.4 其他
Locality：输入数据由GFS管理，3副本，master调度map任务时会考虑数据文件的位置信息；
Backup Tasks：影响一个mapreduce的总执行时间的是“落伍者”，当一个 MapReduce 操作接近完成的时候，master调度备用(backup)任务进程来执行剩下的、处于处理中状态(in-progress)的任务。无论是最初的执行进程、 还是备用(backup)任务进程完成了任务，我们都把这个任务标记成为已经完成。
map全部执行完毕后，才执行reduce？No Reduce calls until all Maps are finished;
load balance : many more tasks than workers, fast workers do more. Task数远多于worker数，性能好的机器执行多任务，性能差的机器执行少任务，从而提高集群的动态的负载均衡能力。
What if the master gives two workers the same Map() task?
perhaps the master incorrectly thinks one worker died.
it will tell Reduce workers about only one of them.What if the master gives two workers the same Reduce() task?
they will both try to write the same output file on GFS!
atomic GFS rename prevents mixing; one complete file will be visible.What if a worker computes incorrect output, due to broken h/w or s/w?
too bad! MR assumes "fail-stop" CPUs and software.
四：总结
MapReduce single-handedly made big cluster computation popular.
- Not the most efficient or flexible.
- Scales well.
- Easy to program -- failures and data movement are hidden.
These were good trade-offs in practice.
[2017.9 梦工厂]
