百万级并发情况的点订单编号的生成方案

项目中要求从一个其实索引开始自增，生成百万级数量数据，要求时间要短，数字的唯一性。
解决方案1 .AtomicLong+CompletableFuture实现秒级数据
2 LongAddr+线程池
经测试发现方案2明显在数据量为100W时数据的重复量为白条这是不可接受的。所以采用方案1

@Test
  public void dotask() {
    long start = System.currentTimeMillis();
    Vector<String> resultList = new Vector<>();
    List<Integer> ids = new ArrayList<>();
    for (int i = 0; i < 100*10000; i++) {
      ids.add(i);
    }
    Set<String> s = Sets.newConcurrentHashSet();
    IterationVariate variate = new IterationVariate(1, 100*10000);
    ConcurrentMap<String, Integer> map = new ConcurrentHashMap<>();
    //RadomVariate variate = new RadomVariate(10,100);
    CompletableFuture[] futures = ids.stream()
        .map(id -> CompletableFuture.supplyAsync(() -> variate.next())
            .whenComplete((res, e) -> {
              //System.out.println("任务"+s+"完成!result="+s+"，异常 e="+e+","+new Date());
              if (map.get(res) != null && map.get(res).intValue() > 0) {
                map.put(res, map.get(res).intValue() + 1);
              } else {
                map.put(res, 1);
              }
            }))
        .toArray(size -> new CompletableFuture[size]);
    //指定该异步方法数组的子任务线程等待类型
    //CompletableFuture.anyOf(futures).join();//anyOf()为任意一个子任务线程执行完毕后返回
    CompletableFuture.allOf(futures).join();
    System.out.println("list=" + resultList + ",耗时=" + (System.currentTimeMillis() - start));
    List<String> list = new ArrayList<>();
    for (String ss : map.keySet()) {
      Integer integer = map.get(ss);
      if (integer != 1) {
        System.out.println("key:" + ss + ",value:" + integer);
      }
      if (integer != null && integer.intValue() > 1) {
        resultList.add(ss);
      }

    }
    System.out.println("resultList.size():" + resultList.size());
    System.out.println("map =" + ",map的Size:" + map.keySet().size());
  }

但是测试后任然发现数据的重复量为10余条后来，采用在生成数据上面加锁实现订单编号的唯一。

@Override
  public String next() {
    System.out.println("线程：" + Thread.currentThread().getName() + ",开始");
    long current;
    synchronized (this) {
      current = start + index++;
      if (current > end) {
        current = start;
        index = 1;
      }
    }
    System.out.println("线程：" + Thread.currentThread().getName() + ",结束");
    return StringUtils.leftPad(Long.toString(current), length, '0');
  }