CompletableFuture的API很丰富，本案例以提交订单为例：
在微服务架构下，划分为很多的模块

• 订单模块
• 购物车模块
• 商品模块
  1、spu模块
  2、sku模块
• 积分模块
• 优惠券模块
• 用户模块
• 附件模块
• 消息模块等等
  在下订单时如何提高这些模块间的执行效率？
  该串行话为并行，CompletableFuture 很好解决了我们的问题，下面我以伪代码的方式进行讲述
  传统模式下：

@RestController
@RefreshScope
public class OrderControler {

    @Autowired
    private SkuService skuService;
    @Autowired
    private SpuService spuService;
    @Autowired
    private PointService pointService;
    @Autowired
    private UserInfoService userService;
    @Autowired
    private MessageService messageService;

    @RequestMapping("/sync/order/sava")
    public String savaOrder(){
        System.out.println("创建订单开始");
        long start = System.currentTimeMillis();
        skuService.querySkuInfo();
        userService.queryUserInfo();
        spuService.querySpuInfo();
        pointService.queryPointInfo();
        messageService.queryMsgInfo();
        long end = System.currentTimeMillis();
        System.out.println("创建订单结束，共花费："+(end-start));
        return "下单花费时间为："+(end-start);
    }
    
}

image.png

使用CompletableFuture之后:

@RestController
@RefreshScope
public class OrderControler {

    @Autowired
    private SkuService skuService;
    @Autowired
    private SpuService spuService;
    @Autowired
    private PointService pointService;
    @Autowired
    private UserInfoService userService;
    @Autowired
    private MessageService messageService;
    

    @RequestMapping("/async/order/sava")
    public String savaOrderAsnc(){
        System.out.println("异步创建订单开始");
        long start = System.currentTimeMillis();
        CompletableFuture<Void> skuServiceFuture = CompletableFuture.runAsync(() -> {
            skuService.querySkuInfo();
        });
        CompletableFuture<Void> userServiceFuture = CompletableFuture.runAsync(() -> {
            userService.queryUserInfo();
        });
        CompletableFuture<Void>  spuServiceFuture = CompletableFuture.runAsync(() -> {
            spuService.querySpuInfo();
        });
        CompletableFuture<Void> pointServiceFuture = CompletableFuture.runAsync(() -> {
            pointService.queryPointInfo();
        });
        CompletableFuture<Void> messageServiceFuture = CompletableFuture.runAsync(() -> {
            messageService.queryMsgInfo();
        });

        CompletableFuture.allOf(
                skuServiceFuture,
                userServiceFuture,
                spuServiceFuture,
                pointServiceFuture,
                messageServiceFuture).join();
        long end = System.currentTimeMillis();
        System.out.println("异步创建订单结束，共花费："+(end-start));
        return "下单花费时间为："+(end-start);
    }

}

image.png

可见性能提升很明显，也进一步说明了木桶短板原理，提升性能就是解决短板问题。