赞
踩
-
- /**
- * 订单号分布式——序列号生成工具
- */
- public class SeqGenerator
- {
- /** 开始时间截 (2017-01-01) */
- private static final long twepoch = 1483200000000L;
-
- /** 时间所占的位数 */
- private static final long timestampBits = 41L;
-
- /** 机器id所占的位数 */
- private static final long workerIdBits = 10L;
-
- /** 支持的最大机器id,结果是31 (这个移位算法可以很快的计算出几位二进制数所能表示的最大十进制数) */
- private static final long maxWorkerId = -1L ^ (-1L << workerIdBits);
-
- /** 序列在id中占的位数 */
- private static final long sequenceBits = 12L;
-
- /** 机器ID向左移12位 */
- private static final long workerIdShift = sequenceBits;
-
- /** 时间截向左移22位(10+12) */
- private static final long timestampLeftShift = sequenceBits + workerIdBits;
-
- /** 生成序列的掩码,这里为4095 (0b111111111111=0xfff=4095) */
- private static final long sequenceMask = -1L ^ (-1L << sequenceBits);
-
- /** 毫秒内序列(0~4095) */
- private static long sequence = 0L;
-
- /** 上次生成ID的时间截 */
- private static long lastTimestamp = -1L;
-
- /**
- * 分布式订单号<br>
- * 为一个64位Long型数字,结构如下(每部分用-分开):<br>
- * 0 - 0000000000 0000000000 0000000000 0000000000 0 - 0000000000 - 000000000000 <br>
- * 1位标识,由于long基本类型在Java中是带符号的,最高位是符号位,正数是0,负数是1,所以id一般是正数,最高位是0<br>
- * 41位时间截(毫秒级),注意,41位时间截不是存储当前时间的时间截,而是存储时间截的差值(当前时间截 - 开始时间截),
- * 这里的的开始时间截,一般是我们的id生成器开始使用的时间,由我们程序来指定的(如下下面程序twepoch属性)。
- * 41位的时间截,可以使用69年,年T = (1L << 41) / (1000L * 60 * 60 * 24 * 365) = 69<br>
- * 10位的数据机器位,可以部署在1024个节点<br>
- * 12位序列,毫秒内的计数,12位的计数顺序号支持每个节点每毫秒(同一机器,同一时间截)产生4096个ID序号
- * @param workerId 工作主机ID(取值:0 ~ 1023)
- */
- public static synchronized long nextId(int workerId)
- {
- if (workerId > maxWorkerId || workerId < 0) {
- throw new IllegalArgumentException(String.format("worker Id can't be greater than %d or less than 0", maxWorkerId));
- }
- long timestamp = timeGen();
- //如果当前时间小于上一次ID生成的时间戳,说明系统时钟回退过这个时候应当抛出异常
- if (timestamp < lastTimestamp) {
- throw new RuntimeException(
- String.format("Clock moved backwards. Refusing to generate id for %d milliseconds", lastTimestamp - timestamp));
- }
- //如果是同一时间生成的,则进行毫秒内序列
- if (lastTimestamp == timestamp) {
- sequence = (sequence + 1) & sequenceMask;
- //毫秒内序列溢出
- if (sequence == 0) {
- //阻塞到下一个毫秒,获得新的时间戳
- timestamp = tilNextMillis(lastTimestamp);
- }
- }
- //时间戳改变,毫秒内序列重置
- else {
- sequence = 0L;
- }
- //上次生成ID的时间截
- lastTimestamp = timestamp;
- //移位并通过或运算拼到一起组成64位的ID
- // System.out.println(String.format("timestamp:%d, workerId:%d, sequence:%d", timestamp - twepoch, workerId, sequence));
- return ((timestamp - twepoch) << timestampLeftShift)
- | (workerId << workerIdShift)
- | sequence;
- }
-
- public static synchronized Long[] parserId(long id)
- {
- long timestamp = id >>> timestampLeftShift;
- long workerId = id << timestampBits + 1;
- workerId = workerId >>> timestampBits + sequenceBits + 1;
- long sequence = id << timestampBits + workerIdBits + 1;
- sequence = sequence >>> timestampBits + workerIdBits + 1;
- // System.out.println(String.format("timestamp:%d, workerId:%d, sequence:%d", timestamp, workerId, sequence));
- return new Long[]{timestamp + twepoch, workerId, sequence};
- }
-
- /**
- * 阻塞到下一个毫秒,直到获得新的时间戳
- * @param lastTimestamp 上次生成ID的时间截
- * @return 当前时间戳
- */
- private static long tilNextMillis(long lastTimestamp) {
- long timestamp = timeGen();
- while (timestamp <= lastTimestamp) {
- timestamp = timeGen();
- }
- return timestamp;
- }
-
- /**
- * 返回以毫秒为单位的当前时间
- * @return 当前时间(毫秒)
- */
- private static long timeGen() {
- return System.currentTimeMillis();
- }
-
- public static void main(String[] args)
- {
- long id = nextId(1024);
- System.out.println(id);
- System.out.println(parserId(id)[1]);
- }
- }
其他文章,相同算法更详细介绍,详见传送门
Copyright © 2003-2013 www.wpsshop.cn 版权所有,并保留所有权利。