Memcached是什麼?
Memcached是由Danga Interactive開發的,高性能的,分佈式的內存對象緩存系統,用於在動態應用中減小數據庫負載,提高訪問速度。html
Memcached能緩存什麼?
經過在內存裏維護一個統一的巨大的hash表,Memcached可以用來存儲各類格式的數據,包括圖像、視頻、文件以及數據庫檢索的結果等。
Memcached快麼?
很是快。Memcached使用了libevent(若是能夠的話,在linux下使用epoll)來均衡任何數量的打開連接,使用非阻塞的網絡I/O,對內部對象實現引用計數(所以,針對多樣的客戶端,對象能夠處在多樣的狀態), 使用本身的頁塊分配器和哈希表, 所以虛擬內存不會產生碎片而且虛擬內存分配的時間複雜度能夠保證爲O(1).。
Memcached的特色?
Memcached的緩存是一種分佈式的,可讓不一樣主機上的多個用戶同時訪問, 所以解決了共享內存只能單機應用的侷限,更不會出現使用數據庫作相似事情的時候,磁盤開銷和阻塞的發生。linux
Memcached的使用須要服務端和客戶端同時合做完成。算法
服務端通常用Danga Interactive開發的:數據庫
Memcached官方站點:http://www.danga.com/memcached/windows
Memcached Win32 1.2.6下載:http://code.jellycan.com/memcached/ 或者 http://code.google.com/p/beitmemcached/(windows)緩存
1.解壓Memcached_1.2.5.zip ,它是memcached的服務器端。
2.把Memcached_1.2.5複製到你指定的作爲緩存服務器的電腦上,好比叫作192.168.0.1。
3.cmd下運行相似命令 'd:/memcached/memcached.exe -d install' 安裝服務器端,這時候它應該會出如今windows服務中
4.cmd下運行相似命令 'd:/memcached/memcached.exe -d start'啓動服務,看服務器進程中是否有memcached進程。
5.確認服務器端口11211是否開放(防火牆設置中),不然其餘機器沒法訪問
6.服務器端這時已經安裝完畢、在其餘機器上測試一下,cmd輸入telnet 192.168.0.1 11211看可否登陸服務器
telnet添加數據到memcached:網絡
add key值 0 時間(秒) 長度(存放多長時間由你指定, 鍵名不能重複,可是值能夠重複app
)例:add name 0 5 5不能夠在add name 0 77 44socket
獲取數據
get key值 【key-val key不能重複,可是val能夠重複】
修改數據
replace key值 0 60 5 【若是key值不存在,則失敗】
set key值 0 60 5 【若是key值不存在,至關於添加,若是存在,則至關於修改.】
刪除數據
delete 鍵值
例:
至此服務端配置完成
客戶端的版本比較多,而且不能互用,由於採用了壓縮機制,日誌等功能,因此在選擇客戶端時要注意這些。
Memcached .NET客戶端:
1).NET memcached client library
下載地址:https://sourceforge.net/projects/memcacheddotnet
2)enyim.com Memcached Client
下載地址:http://www.codeplex.com/EnyimMemcached/
3)Memcached Providers
下載地址:http://www.codeplex.com/memcachedproviders
4) BeIT Memcached
下載地址:http://code.google.com/p/beitmemcached/
我這裏用BeIT,BeITMemcached_source_2008_05_31.zip壓縮包下栽下來,裏面有Sample
class Example { public static void Main(string[] args) { //--------------------- // Setting up a client. //--------------------- Console.Out.WriteLine("Setting up Memcached Client."); MemcachedClient.Setup("MyCache", new string[] { "localhost" }); //It is possible to have several clients with different configurations: //If it is impossible to resolve the hosts, this method will throw an exception. try { MemcachedClient.Setup("MyOtherCache", new string[]{ "server1.example.com:12345", "server2.example.com:12345"}); } catch (Exception e) { Console.WriteLine(e.Message); } //Get the instance we just set up so we can use it. You can either store this reference yourself in //some field, or fetch it every time you need it, it doesn't really matter. MemcachedClient cache = MemcachedClient.GetInstance("MyCache"); //It is also possible to set up clients in the standard config file. Check the section "beitmemcached" //in the App.config file in this project and you will see that a client called "MyConfigFileCache" is defined. MemcachedClient configFileCache = MemcachedClient.GetInstance("MyConfigFileCache"); //Change client settings to values other than the default like this: cache.SendReceiveTimeout = 5000; cache.ConnectTimeout = 5000; cache.MinPoolSize = 1; cache.MaxPoolSize = 5; //---------------- // Using a client. //---------------- //Set some items Console.Out.WriteLine("Storing some items."); cache.Set("mystring", "The quick brown fox jumped over the lazy dog."); cache.Set("myarray", new string[]{"This is the first string.", "This is the second string."}); cache.Set("myinteger", 4711); cache.Set("mydate", new DateTime(2008, 02, 23)); //Use custom hash cache.Set("secondstring", "Flygande b鋍kasiner s鰇a hwila p?mjuka tufvor", 4711); //Get a string string str = cache.Get("mystring") as string; if (str != null) { Console.Out.WriteLine("Fetched item with key: mystring, value: " + str); } //Get an object string[] array = cache.Get("myarray") as string[]; if (array != null) { Console.Out.WriteLine("Fetched items with key: myarray, value 1: " + array[0] + ", value 2: " + array[1]); } //Get several values at once object[] result = cache.Get(new string[]{"myinteger", "mydate"}); if (result[0] != null && result[0] is int) { Console.Out.WriteLine("Fetched item with key: myinteger, value: " + (int)result[0]); } if (result[1] != null && result[1] is DateTime) { Console.Out.WriteLine("Fetched item with key: mydate, value: " + (DateTime)result[1]); } str = cache.Get("secondstring", 4711) as string; if (str != null) { Console.Out.WriteLine("Fetched item with key and custom hash: secondstring, value: " + str); } //Set a counter Console.Out.WriteLine("Setting an item for incrementing and decrementing."); cache.SetCounter("mycounter", 9000); ulong? counter = cache.GetCounter("mycounter"); if (counter.HasValue) { Console.Out.WriteLine("Fetched mycounter, value: " + counter.Value); } //Increment the counter counter = cache.Increment("mycounter", 1); if (counter.HasValue) { Console.Out.WriteLine("Incremented mycounter with 1, new value: " + counter.Value); } //Decrement the counter counter = cache.Decrement("mycounter", 9000); if (counter.HasValue) { Console.Out.WriteLine("Decremented mycounter with 9000, new value: " + counter.Value); } //Append and prepend Console.Out.WriteLine("Storing bar for append/prepend"); cache.Set("foo", "bar"); Console.Out.WriteLine("Appending baz"); cache.Append("foo", " baz"); Console.Out.WriteLine("Prepending foo"); cache.Prepend("foo", "foo "); Console.Out.WriteLine("New value: " + cache.Get("foo")); //Cas cache.Delete("castest"); Console.Out.WriteLine("Trying to CAS non-existant key castest: " + cache.CheckAndSet("castest", "a", 0)); Console.Out.WriteLine("Setting value for key: castest, value: a"); cache.Set("castest", "a"); Console.Out.WriteLine("Trying to CAS key castest with the wrong unique: " + cache.CheckAndSet("castest", "a", 0)); ulong unique; cache.Gets("castest", out unique); Console.Out.WriteLine("Getting cas unique for key castest: " + unique); Console.Out.WriteLine("Trying to CAS again with the above unique: " + cache.CheckAndSet("castest", "b", unique)); string value = cache.Gets("castest", out unique) as string; Console.Out.WriteLine("New value: " + value + ", new unique: " + unique); Console.Out.WriteLine("Displaying the socketpool status:"); foreach (KeyValuePair<string, Dictionary<string, string>> host in cache.Status()) { Console.Out.WriteLine("Host: " + host.Key); foreach (KeyValuePair<string, string> item in host.Value) { Console.Out.WriteLine("\t" + item.Key + ": " + item.Value); } Console.Out.WriteLine(); } Console.Out.WriteLine(); Console.Out.WriteLine("Finished. Press enter to exit."); Console.In.ReadLine(); } }
最先的客戶端是採用取餘的hash算法來查找服務器的,咱們知道以往資料要放到 M 臺服務器上,最簡單的方法就是取餘數 (hash_value % M) 而後放到對應的服務器上,那就是當添加或移除服務器時,緩存重組的代價至關巨大。添加服務器後,餘數就會產生鉅變,這樣就沒法獲取與保存時相同的服務器, 從而影響緩存的命中率。
現在採用了Consistent Hashing算法來改善性能,具體能夠看最後的參考網址。
參考
http://blog.csdn.net/jinjazz/article/details/2664136
http://www.cnblogs.com/dudu/archive/2009/07/19/1526407.html
http://www.cnblogs.com/mecity/archive/2011/06/13/2079548.html
http://www.cnblogs.com/wucg/archive/2011/03/01/1968185.html