滑动窗口计数java实现
滑动窗口计数有很多使用场景,比如说限流防止系统雪崩。相比计数实现,滑动窗口实现会更加平滑,能自动消除毛刺。
概念上可以参考TCP的滑窗算法,可以看一下这篇文章(http://go12345.iteye.com/blog/1744728)。在实现上,滑动窗口算法需要循环队列和线程安全保障。
下面的实现有几个点
1, 支持滑窗大小运行时动态调整
2, 基于 java8 编译器
3, DEMO实现只支持一个窗口对象,如果要支持多个,需要修改 SlotBaseCounter 类
- package slidingwindow;
- import java.util.Arrays;
- import java.util.concurrent.atomic.AtomicInteger;
- /**
- * Created by admin on 2016/02/20.
- */
- public class SlotBaseCounter {
- private int slotSize;
- private AtomicInteger[] slotCounter;
- public SlotBaseCounter(int slotSize) {
- slotSize = slotSize < 1 ? 1 : slotSize;
- this.slotSize = slotSize;
- this.slotCounter = new AtomicInteger[slotSize];
- for (int i = 0; i < this.slotSize; i++) {
- slotCounter[i] = new AtomicInteger(0);
- }
- }
- public void increaseSlot(int slotSize) {
- slotCounter[slotSize].incrementAndGet();
- }
- public void wipeSlot(int slotSize) {
- slotCounter[slotSize].set(0);
- }
- public int totalCount() {
- return Arrays.stream(slotCounter).mapToInt(slotCounter -> slotCounter.get()).sum();
- }
- @Override
- public String toString() {
- return Arrays.toString(slotCounter);
- }
- }
- package slidingwindow;
- /**
- * Created by admin on 2016/02/20.
- */
- public class SlidingWindowCounter {
- private volatile SlotBaseCounter slotBaseCounter;
- private volatile int windowSize;
- private volatile int head;
- public SlidingWindowCounter(int windowSize) {
- resizeWindow(windowSize);
- }
- public synchronized void resizeWindow(int windowSize) {
- this.windowSize = windowSize;
- this.slotBaseCounter = new SlotBaseCounter(windowSize);
- this.head = 0;
- }
- public void increase() {
- slotBaseCounter.increaseSlot(head);
- }
- public int totalAndAdvance() {
- int total = totalCount();
- advance();
- return total;
- }
- public void advance() {
- int tail = (head + 1) % windowSize;
- slotBaseCounter.wipeSlot(tail);
- head = tail;
- }
- public int totalCount() {
- return slotBaseCounter.totalCount();
- }
- @Override
- public String toString() {
- return "total = " + totalCount() + " head = " + head + " >> " + slotBaseCounter;
- }
- }
- package slidingwindow;
- import java.util.concurrent.TimeUnit;
- /**
- * Created by admin on 2016/02/20.
- */
- public class Loops {
- public static void dieLoop(Runnable runnable) {
- while (true) {
- run(runnable);
- }
- }
- public static void rateLoop(Runnable runnable, int mills) {
- while (true) {
- try {
- TimeUnit.MILLISECONDS.sleep(mills);
- } catch (InterruptedException e) {
- }
- run(runnable);
- }
- }
- public static void fixLoop(Runnable runnable, int loop) {
- for (int i = 0; i < loop; i++) {
- run(runnable);
- }
- }
- private static void run(Runnable runnable) {
- try {
- runnable.run();
- } catch (Exception e) {
- e.printStackTrace();
- }
- }
- }
- package slidingwindow;
- import org.junit.Test;
- import java.io.IOException;
- import java.util.Random;
- import java.util.Scanner;
- import java.util.concurrent.ExecutorService;
- import java.util.concurrent.Executors;
- import java.util.concurrent.ScheduledExecutorService;
- import java.util.concurrent.TimeUnit;
- /**
- * Created by admin on 2016/02/20.
- */
- public class SlidingWindowCounterTest {
- private ExecutorService es = Executors.newCachedThreadPool();
- private ScheduledExecutorService ses = Executors.newSingleThreadScheduledExecutor();
- @Test
- public void testNWindow() throws IOException {
- SlidingWindowCounter swc = new SlidingWindowCounter(3);
- ses.scheduleAtFixedRate(() -> {
- Loops.fixLoop(swc::increase, new Random().nextInt(10));
- }, 10, 2, TimeUnit.MILLISECONDS);
- ses.scheduleAtFixedRate(() -> {
- System.out.println(swc);
- swc.advance();
- }, 1, 1, TimeUnit.SECONDS);
- ses.scheduleAtFixedRate(() -> {
- swc.resizeWindow(new Random().nextInt(10));
- }, 1, 10, TimeUnit.SECONDS);
- System.in.read();
- }
- @Test
- public void test1Window() {
- SlidingWindowCounter swc = new SlidingWindowCounter(1);
- System.out.println(swc);
- swc.increase();
- swc.increase();
- System.out.println(swc);
- swc.advance();
- System.out.println(swc);
- swc.increase();
- swc.increase();
- System.out.println(swc);
- }
- @Test
- public void test3Window() {
- SlidingWindowCounter swc = new SlidingWindowCounter(3);
- System.out.println(swc);
- swc.increase();
- System.out.println(swc);
- swc.advance();
- System.out.println(swc);
- swc.increase();
- swc.increase();
- System.out.println(swc);
- swc.advance();
- System.out.println(swc);
- swc.increase();
- swc.increase();
- swc.increase();
- System.out.println(swc);
- swc.advance();
- System.out.println(swc);
- swc.increase();
- swc.increase();
- swc.increase();
- swc.increase();
- System.out.println(swc);
- swc.advance();
- System.out.println(swc);
- }
- }
这是部分测试结果输出:
total = 2245 head = 0 >> [2245, 0, 0, 0, 0, 0]
total = 4561 head = 1 >> [2245, 2316, 0, 0, 0, 0]
total = 6840 head = 2 >> [2245, 2316, 2279, 0, 0, 0]
total = 8994 head = 3 >> [2245, 2316, 2279, 2154, 0, 0]
total = 11219 head = 4 >> [2245, 2316, 2279, 2154, 2225, 0]
total = 13508 head = 5 >> [2245, 2316, 2279, 2154, 2225, 2289]
total = 13602 head = 0 >> [2339, 2316, 2279, 2154, 2225, 2289]
total = 13465 head = 1 >> [2339, 2179, 2279, 2154, 2225, 2289]
total = 13474 head = 2 >> [2339, 2179, 2288, 2154, 2225, 2289]
total = 13551 head = 3 >> [2339, 2179, 2288, 2231, 2225, 2289]
total = 2192 head = 0 >> [2192]
total = 2207 head = 0 >> [2207]
total = 2291 head = 0 >> [2291]
total = 2257 head = 0 >> [2257]
total = 2250 head = 0 >> [2250]
total = 2201 head = 0 >> [2201]
total = 2299 head = 0 >> [2299]
total = 2223 head = 0 >> [2223]
total = 2190 head = 0 >> [2190]
total = 2306 head = 0 >> [2306]
total = 2290 head = 0 >> [2290, 0, 0, 0, 0, 0, 0, 0, 0]
total = 4474 head = 1 >> [2290, 2184, 0, 0, 0, 0, 0, 0, 0]
total = 6632 head = 2 >> [2290, 2184, 2158, 0, 0, 0, 0, 0, 0]
total = 8744 head = 3 >> [2290, 2184, 2158, 2112, 0, 0, 0, 0, 0]
total = 11008 head = 4 >> [2290, 2184, 2158, 2112, 2264, 0, 0, 0, 0]
total = 13277 head = 5 >> [2290, 2184, 2158, 2112, 2264, 2269, 0, 0, 0]
total = 15446 head = 6 >> [2290, 2184, 2158, 2112, 2264, 2269, 2169, 0, 0]
total = 17617 head = 7 >> [2290, 2184, 2158, 2112, 2264, 2269, 2169, 2171, 0]
total = 19749 head = 8 >> [2290, 2184, 2158, 2112, 2264, 2269, 2169, 2171, 2132]
total = 19608 head = 0 >> [2149, 2184, 2158, 2112, 2264, 2269, 2169, 2171, 2132]
total = 2256 head = 0 >> [2256, 0, 0, 0]
total = 4624 head = 1 >> [2256, 2368, 0, 0]
total = 6811 head = 2 >> [2256, 2368, 2187, 0]
total = 8973 head = 3 >> [2256, 2368, 2187, 2162]
total = 8934 head = 0 >> [2217, 2368, 2187, 2162]
total = 8798 head = 1 >> [2217, 2232, 2187, 2162]
total = 8912 head = 2 >> [2217, 2232, 2301, 2162]
total = 8940 head = 3 >> [2217, 2232, 2301, 2190]
total = 8987 head = 0 >> [2264, 2232, 2301, 2190]
total = 9049 head = 1 >> [2264, 2294, 2301, 2190]
total = 2220 head = 0 >> [2220, 0, 0, 0, 0, 0, 0]
total = 4477 head = 1 >> [2220, 2257, 0, 0, 0, 0, 0]
total = 6718 head = 2 >> [2220, 2257, 2241, 0, 0, 0, 0]
total = 8939 head = 3 >> [2220, 2257, 2241, 2221, 0, 0, 0]
total = 11174 head = 4 >> [2220, 2257, 2241, 2221, 2235, 0, 0]
total = 13420 head = 5 >> [2220, 2257, 2241, 2221, 2235, 2246, 0]
total = 15673 head = 6 >> [2220, 2257, 2241, 2221, 2235, 2246, 2253]
total = 15779 head = 0 >> [2326, 2257, 2241, 2221, 2235, 2246, 2253]
total = 15796 head = 1 >> [2326, 2274, 2241, 2221, 2235, 2246, 2253]
total = 15802 head = 2 >> [2326, 2274, 2247, 2221, 2235, 2246, 2253]
滑动窗口计数java实现的更多相关文章
- Storm 实现滑动窗口计数和TopN排序
计算top N words的topology, 用于比如trending topics or trending images on Twitter. 实现了滑动窗口计数和TopN排序, 比较有意思, ...
- 【Java】 剑指offer(59-1) 滑动窗口的最大值
本文参考自<剑指offer>一书,代码采用Java语言. 更多:<剑指Offer>Java实现合集 题目 给定一个数组和滑动窗口的大小,请找出所有滑动窗口里的最大值.例 ...
- Leetcode 239题 滑动窗口最大值(Sliding Window Maximum) Java语言求解
题目链接 https://leetcode-cn.com/problems/sliding-window-maximum/ 题目内容 给定一个数组 nums,有一个大小为 k 的滑动窗口从数组的最左侧 ...
- Java实现 LeetCode 480 滑动窗口中位数
480. 滑动窗口中位数 中位数是有序序列最中间的那个数.如果序列的大小是偶数,则没有最中间的数:此时中位数是最中间的两个数的平均数. 例如: [2,3,4],中位数是 3 [2,3],中位数是 (2 ...
- Java实现 LeetCode 239 滑动窗口最大值
239. 滑动窗口最大值 给定一个数组 nums,有一个大小为 k 的滑动窗口从数组的最左侧移动到数组的最右侧.你只可以看到在滑动窗口内的 k 个数字.滑动窗口每次只向右移动一位. 返回滑动窗口中的最 ...
- Sentinel源码分析-滑动窗口统计原理
滑动窗口技术是Sentinel比较关键的核心技术,主要用于数据统计 通过分析StatisticSlot来慢慢引出这个概念 @Override public void entry(Context con ...
- uva 1606 amphiphilic carbon molecules【把缩写写出来,有惊喜】(滑动窗口)——yhx
Shanghai Hypercomputers, the world's largest computer chip manufacturer, has invented a new classof ...
- hdu-5497 Inversion(滑动窗口+树状数组)
题目链接: Inversion Time Limit: 6000/3000 MS (Java/Others) Memory Limit: 65536/65536 K (Java/Others)T ...
- storm 1.0版本滑动窗口的实现及原理
滑动窗口在监控和统计应用的场景比较广泛,比如每隔一段时间(10s)统计最近30s的请求量或者异常次数,根据请求或者异常次数采取相应措施.在storm1.0版本之前,没有提供关于滑动窗口的实现,需要开发 ...
随机推荐
- How to copy projects into workspace of eclipse after importing the project?
在eclipse中如果已经导入了一个别处的项目但导入时没有选“copy projects into workspace”怎么办? 答案是删掉该项目重新导入... http://stackoverflo ...
- Jackson的高级应用(转)
Jackson 是当前用的比较广泛的,用来序列化和反序列化 JSON 的 Java 的开源框架.Jackson 社 区相对比较活跃,更新速度也比较快, 从 Github 中的统计来看,Jackson ...
- RTM CU CTP
.RTM: Released To Manufacturing,也就是新产品的正式发布版本,比如SQL 2008 RTM.其后的补丁都是基于这个之上的. .CU: Cumulative Update, ...
- iptables 要点总结
http://jiayu0x.com/2014/12/02/iptables-essential-summary/
- delphi 浮点数转换成十六进制字符串的方法
我们在研究封包技术时,经常会碰到将浮点数转换成十六进制形式.比如在游戏中人物的座标,经常就用浮点数来表示.怎么将浮点数转换成十六进制字符串形式呢?下面我将写出其在DELPHI中的方法. 先 ...
- 【Linux】CentOS7上的一些操作小方法
1.在文件夹目录下删除文件 点击文件,按Delete键删除,就可以把文件删除到回收站中. 2.更改命令窗口的样式 打开命令窗口--->右键---->配置文件---->配置文件首选项
- Oracle数据库实现获取前几条数据的方法
如何在Oracle数据库中实现获取前几条数据的方法呢?就是类似SQL语句中的SELECT TOP N的方法.本文将告诉您答案,举例说明了哟! 1.在Oracle中实现SELECT TOP N : ...
- Git历险记(二)——Git的安装和配置
各位同学,上回Git历险记(一)讲了一个 “hello Git” 的小故事.有的同学可能是玩过了其它分布式版本控制系统(DVCS),看完之后就触类旁通对Git就了然于胸了:也有的同学可能还如我当初入手 ...
- Java高级特性—反射和动态代理
1).反射 通过反射的方式可以获取class对象中的属性.方法.构造函数等,一下是实例: 2).动态代理 使用场景: 在之前的代码调用阶段,我们用action调用service的方法实现业务即可. 由 ...
- WEB接口测试之Jmeter接口测试自动化 (二)(数据分离)
转载: http://www.cnblogs.com/chengtch/p/6105231.html 通过逐个录入的方式,好不容易将需要测试几十个接口的300多个测试用例录入sampler-ht ...