Caffeine,HashMap,Redis性能对比压测

服务端是Java 单机，Redis在远程服务器上，服务启动时往缓存里放入5000个字符串。

    private final RedisUtil redisUtil;

    private final HashMap<String, String> hashMap = new HashMap<>();

    private final Cache<String, String> caffeineCache = Caffeine.newBuilder().expireAfterWrite(200000, TimeUnit.SECONDS)
            .maximumSize(10000).build();

    @PostConstruct
    private void init() {
        for (int i = 0; i < 5000; i++) {
            String kv = "MISTRA" + i;
            redisUtil.set(kv, kv);
            hashMap.put(kv, kv);
            caffeineCache.put(kv, kv);
        }
        redisUtil.set("MISTRA", "mistra#");
        hashMap.put("MISTRA", "mistra#");
        caffeineCache.put("MISTRA", "mistra#");
    }

    @PostMapping("/test-redis")
    public String test() {
        return redisUtil.get("MISTRA");
    }

    @PostMapping("/test-hashMap")
    public String testMap() {
        return hashMap.get("MISTRA");
    }

    @PostMapping("/test-caffeine")
    public String testCaffeine() {
        return caffeineCache.getIfPresent("MISTRA");
    }

JMeter线程组设置：1秒内启动1100个线程，每个线程循环请求500次

Caffeine QPS:30012

HashMap QPS:30545

Redis QPS:1359
因为Redis要花去网络开销，所以Java单机QPS受限

---