Topic of today's post will be dedicated to running in a cloud of distributed caching mechanism called Memcached. Why is proper Memcached? .
Windows Azure Caching. ), Which is an excellent choice for developers seeking to add a layer of caching in your Windows Azure application. Nevertheless, there are situations that require the use of an alternative mechanism, namely Memcached.
For example, I believe that the application is fully justified memkesha in the presence of our web roles additional RAM. In this case, we are able to raise an additional layer of distributed caching in memory without the need to run more virtual machines (and hence no extra cost ). Yes. Windows Azure Saching. has in its arsenal a powerful thing as a local cache, but the use of this feature is also required all remote cache, and also the local cache is available to only one instance ( so it is actually called a local ).
Another reason why people choose memkesh - a possibility of full manual control of the behavior of cache memory. It's certainly not for the faint of heart, but it is a good option for those people who are experts in tuning memkesha.
How does it work?. So, we decided nevertheless to use as a base layer memkesh distributed caching. To do this, we already have ready details of: the server side and client are available as packages NuGet. The server part is relatively simple - it just runs and opens memkesh daemon listening for incoming connections on a port. The client part is a little more complicated for the simple reason that it must fulfill two basic requirements:.
use. grease hashing. That minimize disruptions to add and remove servers.
Automatically respond to adding and removing servers from the cluster (during zooming, upgrades and downs ).
The first requirement is satisfied by using.
Enyim Memcached Client. , Which by default uses a hashing grease. The second requirement is achieved by expanding.
Enyim. by implementing custom interface.
IServerPool. under the title.
WindowsAzureServerPool. Regularly checks the cluster ID for added and removed servers and in the case of changing the number of servers again reconfigures memkesh client. It is important here is that the new server is activated only when he becomes able to accept incoming connections, but not at the time when he started.
Installing and configuring the server side. Sam memkesh server, we are able to run in the web or worker role. In heavy duty applications, I would recommend running the server in the context of an individual worker- role. In other cases, a better option to start the server option in the context of the role of running the web-. To install and run the server must perform the following steps:.
With NuGet install. server. ( done in the Package Manager using the command. install-package WazMemcachedServer. ). After this, the project will be connected all the necessary binaries and a small helper- class.
Add an internal TCP access point through which will be listening for incoming connections. ( this can be done in Visual Studio by double clicking on the role and select the category ...
Add code to run in our memkesha WebRole. cs or WorkerRole. cs.
Process proc;.
public override void Run ().
{.
proc. WaitForExit ();.
}.
public override bool OnStart ().
{.
proc = WindowsAzureMemcachedHelpers. StartMemcached ('Memcached', 512);.
return base. OnStart ();.
}.
The first parameter - the name of the access point. we created in the 2nd paragraph, the second parameter - the size of a redundant server memory. Pay attention to the method of.
Run. - He listens to memcached- process ( it may be forever:)) for output, so if our memkesh server for some reason zavernetsya, infrastructure, Windows Azure will automatically restart all of us. Once all the items we have successfully implemented our server will be able to accept connections.
Installing and configuring the client side. To use Memcached cluster of our program, you must configure the client that knows how to find copies of the server, even if they are added and removed by scaling or upgrades. Installing the client is not has high complexity, in order to do this perform the following steps:.
With NuGet install. client libraries. ( done in the Package Manager using the command. install-package. WazMemcachedClient. ). This will add a small subset of classes that extend the client. Enyim for Memcached.
Create an instance of the code of the client ( as an alternative storage facility, you can use a static variable ) capable of interfacing with the server.
static MemcachedClient client = WindowsAzureMemcachedHelpers. CreateDefaultClient ('WorkerRole', 'Memcached');. The first parameter - the name of the role which launched memkesh and the second parameter - the name of the internal access point through which the listening connections. Another convenient way to initialize a client - an application is launched.
Application ['memcache'] = WindowsAzureMemcachedHelpers. CreateDefaultClient ('WorkerRole', 'Memcached');. Once our client has been successfully initialized, we get access to the full range of operations available. For example:.
string value = client. Get (key) as string;.
if (value == null).
{.
value = FetchFromStorage (key);.
client. Store (StoreMode. Set, key, value);.
}.
return value;.
So, we met with you so you can start and how to work with the Memcached in Windows Azure. To do this we need two basic package ( server and client side ):.
Blogpost sources:. http://blog. smarx. com / posts / memcached-in-windows-azure. ,. http://www. davidaiken. com/2011/01/11/windows-azure-memcached-plugin /.
