context: a key-value store = Map ▪ Where do we use this caching? • Any method that always returns the same result for the same argument(s) • This method could do anything • Calculate data on the fly • Execute a database query • Request data via RMI, JMS, a web-service ... • A unique key must be generated from the arguments • That's the cache key
caching to Spring beans (AOP) • Define one or more caches in Spring configuration • Mark methods cacheable • Indicate caching key(s) • Name of cache to use (multiple caches supported) • Use annotations or XML @Cacheable public Country[] loadAllCountries() { … } Spring Application In-memory cache Spring 3.1
• its result is stored in a cache • subsequent invocations (with the same arguments) • fetch data from cache using key, method not executed ▪ @Cacheable attributes • value: name of cache to use • key: the key for each cached data-item • Uses SpEL and argument(s) of method @Cacheable(value="books", key="#refId.toUpperCase()") public Book findBook(String refId) {...}
findBook(String title, boolean checkWarehouse); @Cacheable(value="books", key="#author.name") public Book findBook2(Author author, boolean checkWarehouse); @Cacheable(value="books", key="T(example.KeyGen).hash(#author)") public Book findBook3(Author author, boolean checkWarehouse); @CacheEvict(value="books") public Book loadBooks(); @Configuration @EnableCaching public class MyConfig { @Bean public class BookService bookService() { } @Configuration @EnableCaching public class MyConfig { @Bean public class BookService bookService() { } clear cache before method invoked use object property Only cache if condition true custom key generator Use 'books' cache
Redis as a caching layer for the entire network. Kyle recently mentioned the specs of the machine in passing. Prior to the NY move our Redis instance was in a virtual machine on a (lightly loaded) web tier machine. In our (admittedly limited) experience, Redis is so fast that the slowest part of a cache lookup is the time spent reading and writing bytes to the network. This is not surprising, really, if you think about it. http://meta.stackoverflow.com/questions/69164/does-stack-overflow-use-caching-and-if-so-how/69172#69172
activity feed, sessions system, and other services, it runs on several Quadruple Extra-Large Memory instances. Occasionally shard across instances. Redis runs in a master-replica setup. Replicas constantly save to disk. EBS snapshots backup the DB dumps. Dumping on the DB on the master was too taxing. http://highscalability.com/blog/2012/4/9/the-instagram-architecture-facebook-bought-for-a-cool-billio.html
If key already holds a value, it is overwritten, regardless of its type. Get the value of key. If the key does not exist the special value nil is returned. An error is returned if the value stored at key is not a string, If the values of your keys are integers, then you can use DECR or DECRBY to decrement and INCR or INCRBY to increment them. These operations are useful in scenarios in which you want to maintain a count of something – say , the number of hits for a web page redis> SET hello "world" OK > GET hello "world" redis> INCR visits (integer) 1 > INCR visits (integer) 2 > DECR visits (integer) 1 > INCRBY visits 41 (integer) 42
pair against a hash. This option can be useful in scenarios in which you want to save an object with several properties, as in this example: redis> HSET user:1 firstname salvatore (integer) 1 > HSET user:1 lasname sanfilipo (integer) 1 > HGET user:1 firstname "salvatore" > HGETALL user:1 1) "firstname" 2) "salvatore" 3) "lasname" 4) "sanfilipo"
"2010-09-07", type: "cat"}' > SET pet:2 '{name: "Basil", bdate: "2012-08-06", type: "hamster"}' > LPUSH visits pet:1 > LPUSH visits pet:2 > LPUSH visits pet:1 # Read > LRANGE visits 0 -1 > MGET pet:1 pet:2 1) "{name: "Leo", bdate: "2010-09-07", type: "cat"}" 2) "{name: "Basil", bdate: "2012-08-06", type: "hamster"}" Returns the values of all specified keys. For every key that does not hold a string value or does not exist, the special value nil is returned. Because of this, the operation never fails.
file Replication Master/Slave #redis.conf 127.0.0.1 6379 Single Threaded/ single Core utiliztion Shared Memory, Mqs, Semaphores, PubSub Data structures Strings, Lists, Hashes, Sets, Sorted sets. Caching, Expiration, Eviction Policy EXPIRE key seconds Set a key's time to live in seconds EXPIREAT key timestamp Set the expiration for a key as a UNIX timestamp Redis commands give time complexity in big-O notation O(1) < O(log(n)) < O(n) < O(n^2)
</dependencies> Spring Data Redis, part of the larger Spring Data family, provides easy configuration and access to Redis from Spring applications. It offers both low-level and high-level abstractions for interacting with the store, freeing the user from infrastructural concerns.
private RedisTemplate<String, String> template; // inject the template as ListOperations // can also inject as Value, Set, ZSet, and HashOperations @Resource(name="redisTemplate") private ListOperations<String, String> listOps; public void addLink(String userId, URL url) { listOps.leftPush(userId, url.toExternalForm()); // or use template directly redisTemplate.boundListOps(userId).leftPush(url.toExternalForm()); } }
:Modern Approach to Data Mgmt § Virtual data ‘fabric’ manages data in-memory using low cost x-86 machines for elastic scalability § Deploy locally or as high speed data backbone across global regions and private/ public clouds § Use as operational data store or to complement existing RDBMS File Systems Databases Mainframes / other GemFire/SQLFire “Distributed caching platforms (DCPs) have proved to be key enablers for mainframe and DBMS offloading, performance and scalability boosting in Web commerce applications, high-performance publish and subscribe, cloud and software as a service (SaaS), and other scenarios.” – Gartner Emerging Technology Analysis: Distributed Caching Platforms Enable Innovative Scenarios; November 2, 2010
Idw Customer Id1 Order Idx Customer Id2 Order Idy Customer Id3 Order Idz Customer Idn ProductA ProductB ProductC ProductA ProductB ProductC Master/Reference Data Replicated Active Data Partitioned & Colocated
Idw Customer Id1 Order Idx Customer Id2 Order Idy Customer Id3 Order Idz Customer Idn ProductA ProductB ProductC ProductA ProductB ProductC Master/Reference Data Replicated Active Data Partitioned & Colocated
michaelisvy
hiroya0601
goyoki
javiergs
riofujimon
phigasui
sakahukamaki
konokenj
techouse
ryounasso
d_endo
ivargrimstad
tammyeverts
samlambert
addyosmani
eileencodes
rasmusluckow
phodgson
tenderlove
smashingmag
nonsquared
moore
jponch
andyhume
