Redis in Practice: Scenarios, Performance and Practice with PHP
Download as PPT, PDF3 likes1,666 views
Knowledge sharing about Redis, mainly focusing on: Why to use Redis? Comparison of some in-memory storages and their scenarios How to make Redis faster? Consider time complexity, communication latency and serialization Practice of replication and sentinel in PHP
![• Session
[Servlet API]
HttpSession session = req.getSession(); // got by session id
Object oldValue = session.getAttribute(key);
session.setAttribute(key, newValue);
[Implementation]
Scenarios
Java Redis
HashMap<String, HashMap> Hash
map.get(id).get(key); HGET id key
map.get(id).put(key, value); HSET id key value](https://image.slidesharecdn.com/redisinpractice-150919010553-lva1-app6892/85/Redis-in-Practice-Scenarios-Performance-and-Practice-with-PHP-6-320.jpg)
![Scenarios
• Session Timeout
[Servlet API]
session.set(key, new HttpSessionBindingListener() {
public void valueUnbound(HttpSessionBindingEvent event) {
// Trigger Timeout Event
}
});
[Implementation]
Java Redis
LinkedHashMap Hash + Sorted Set (ZSet)
Iterator i = map.values().iterator();
while (i.hasNext() &&
isTimeout(i.next())) {
// Trigger Timeout Event
i.remove();
}
ZRANGEBYSCORE zset 0 now
// Trigger Timeout Event
HDEL id, id, ...
ZREMRANGEBYSCORE zset 0 now](https://image.slidesharecdn.com/redisinpractice-150919010553-lva1-app6892/85/Redis-in-Practice-Scenarios-Performance-and-Practice-with-PHP-7-320.jpg)
![Communication Latency
• Massive (Multi-Row) Insertion
[SQL]
× INSERT INTO cache (key, value) VALUES ("key1",
"value1");
INSERT INTO cache (key, value) VALUES ("key2",
"value2");
...
√ INSERT INTO cache (key, value) VALUES ("key1",
"value1"),
("key2", "value2"), ("key3", "value3"), ...
[Redis]
× HSET cache key1 value1
HSET cache key2 value2
...
√ HMSET cache key1 value1 key2 value2 ...](https://image.slidesharecdn.com/redisinpractice-150919010553-lva1-app6892/85/Redis-in-Practice-Scenarios-Performance-and-Practice-with-PHP-11-320.jpg)
![Communication Latency
• Pipeline (Batch)
[SQL in Java]
try (Statement stmt = conn.createStatement()) {
stmt.addBatch("INSERT INTO cache1 ...");
stmt.addBatch("INSERT INTO cache2 ...");
stmt.executeBatch();
}
[Redis in Java]
Pipeline pl = jedis.pipelined();
pl.hset("cache1", "key1", "value1");
pl.hset("cache2", "key2", "value2");
pl.sync();](https://image.slidesharecdn.com/redisinpractice-150919010553-lva1-app6892/85/Redis-in-Practice-Scenarios-Performance-and-Practice-with-PHP-13-320.jpg)
![Communication Latency
• Scripting
[Java]
int balance = Integer.parseInt(jedis.get("balance"));
if (balance > 100) {
jedis.set("balance", Integer.toString(balance - 100));
}
[Lua]
local i = tonumber(redis.call('get', 'balance'))
if i > 100 then
redis.call('set', 'balance', tostring(i - 100))
end](https://image.slidesharecdn.com/redisinpractice-150919010553-lva1-app6892/85/Redis-in-Practice-Scenarios-Performance-and-Practice-with-PHP-14-320.jpg)
Ad
More Related Content
What's hot (20)
Viewers also liked (8)
Ad
Similar to Redis in Practice: Scenarios, Performance and Practice with PHP (20)
Ad
Recently uploaded (19)
Redis in Practice: Scenarios, Performance and Practice with PHP
- 1. Redis in Practice • Why Redis • How to make Redis Faster • Replication and Sentinel Chen Huang c.huang@*****.com
- 2. Content • Why Redis Comparison of In-Memory Storages Scenarios Sharding (Parallel Scaling) • How to make Redis Faster • Replication and Sentinel
- 3. Comparison of In-Memory Storages Redis Memcached MemSQL Data Structure Key-Value, List, Hash, Set, Z-Set Key-Value Relational Concurrency Poor Good Good Sharding Not Generally Available Yes Yes Transaction Optimistic Lock CAS No Scripting Lua No No Replication Yes Repcached Yes Persistance Yes MemcacheDB Yes High Availability Redis Sentinel ? ?
- 4. Scenarios • Cache Java: Map cache = new LinkedHashMap() { protected boolean removeEldestEntry(Map.Entry eldest) { return isTimeout(eldest); } } Redis: SETEX key timeout value
- 5. Scenarios • List / Queue Java Redis LinkedList / ArrayDeque List list.offerFirst(element) LPUSH list element list.offerLast(element) RPUSH list element list.pollFirst() LPOP list list.pollLast() RPOP list
- 6. • Session [Servlet API] HttpSession session = req.getSession(); // got by session id Object oldValue = session.getAttribute(key); session.setAttribute(key, newValue); [Implementation] Scenarios Java Redis HashMap<String, HashMap> Hash map.get(id).get(key); HGET id key map.get(id).put(key, value); HSET id key value
- 7. Scenarios • Session Timeout [Servlet API] session.set(key, new HttpSessionBindingListener() { public void valueUnbound(HttpSessionBindingEvent event) { // Trigger Timeout Event } }); [Implementation] Java Redis LinkedHashMap Hash + Sorted Set (ZSet) Iterator i = map.values().iterator(); while (i.hasNext() && isTimeout(i.next())) { // Trigger Timeout Event i.remove(); } ZRANGEBYSCORE zset 0 now // Trigger Timeout Event HDEL id, id, ... ZREMRANGEBYSCORE zset 0 now
- 8. Sharding (Parallel Scaling) • Redis is Single-Threaded • Multiple Redis Instances in one Server • Partition by Hash of Keys
- 9. Content • Why Redis • How to make Redis Faster Time Complexity Communication Latency Serialization • Replication and Sentinel
- 10. Time Complexity See Redis Documentation for more Details Structure Operation Complexity Hash HGET id key Key-Value GET id Key-Value KEYS prefix* List LPUSH/LPOP List LINDEX/LSET Sorted Set ZADD Sorted Set ZRANGEBYSCORE O(1) O(1) O(n) O(1) O(n) O(log(n)) O(log(n) +m)
- 11. Communication Latency • Massive (Multi-Row) Insertion [SQL] × INSERT INTO cache (key, value) VALUES ("key1", "value1"); INSERT INTO cache (key, value) VALUES ("key2", "value2"); ... √ INSERT INTO cache (key, value) VALUES ("key1", "value1"), ("key2", "value2"), ("key3", "value3"), ... [Redis] × HSET cache key1 value1 HSET cache key2 value2 ... √ HMSET cache key1 value1 key2 value2 ...
- 12. Communication Latency • Pipeline (Batch)
- 13. Communication Latency • Pipeline (Batch) [SQL in Java] try (Statement stmt = conn.createStatement()) { stmt.addBatch("INSERT INTO cache1 ..."); stmt.addBatch("INSERT INTO cache2 ..."); stmt.executeBatch(); } [Redis in Java] Pipeline pl = jedis.pipelined(); pl.hset("cache1", "key1", "value1"); pl.hset("cache2", "key2", "value2"); pl.sync();
- 14. Communication Latency • Scripting [Java] int balance = Integer.parseInt(jedis.get("balance")); if (balance > 100) { jedis.set("balance", Integer.toString(balance - 100)); } [Lua] local i = tonumber(redis.call('get', 'balance')) if i > 100 then redis.call('set', 'balance', tostring(i - 100)) end
- 15. Communication Latency • SQL Efficiency Multi-Row Insertion > Batch > Stored Procedure • Redis Efficiency Massive Insertion > Pipeline > Scripting
- 16. Serialization
- 17. Serialization
- 18. Serialization • Kryo vs. Protobuf
- 19. Content • Why Redis • How to make Redis Faster • Replication and Sentinel Replication Sentinel Practice with PHP
- 20. Replication Master Slave-1 Slave-2 Slave-3 Web Server Write Sync to Sync to Sync to Web Server Read Read Read
- 22. Practice with PHP Shard 1 Shard 2 Sentinels Shard 3 Shard 4 Web Server Config Center Write Read Scheduled Config Sync Web Server Web Server redis.config.php redis.config.php redis.config.php redis.config.php Scheduled Config Sync