There is a great news for people that use Azure SQL Database. Elastic scale is available in preview phase. I expecting this feature from some time and now I’m happy that is available to us.
The biggest advantage of a cloud solution is scalability and pay as you go. You pay a resource only when you need it without paying the period of time when you don’t use it. Over this features, a service like automatically scaling can be added very easily. This is applicable when we are talking about web sites, web/worker roles where things are not so complicated.
To be able to scale the data-tier or an application you need to be able to use sharding (joining multiple resources, in our case splitting a big database on multiple databases). This feature was enabled on Microsoft Azure in a very smart way and without adding additional costs to the end user. You will pay the Azure SQL instances that you use. For each shard you will need to use a new instance of Azure SQL Database.
Elastic scale is enabled from a simple template that is created using Visual Studio. From this project, we can control and manage the shards, shard mapping, query multiple shards and adding additional one if needed.
When we should use sharding?
Key features available in this moment
How to get a db connection
To get a connection to a specific database you can specify the ID of the entity. In this way you will get a connection directly to the database where your entity is stored.
Conclusion
This is a great feature, that can save your life. Now, sharding can be configured very easily by anyone.
The biggest advantage of a cloud solution is scalability and pay as you go. You pay a resource only when you need it without paying the period of time when you don’t use it. Over this features, a service like automatically scaling can be added very easily. This is applicable when we are talking about web sites, web/worker roles where things are not so complicated.
Elastic scale is enabled from a simple template that is created using Visual Studio. From this project, we can control and manage the shards, shard mapping, query multiple shards and adding additional one if needed.
When we should use sharding?
- The total data that we want to store into one database is too big.
- A part of our database should reside in a specific geographical location (a part in Europe, a part in USA)
- The number of transactions that are executed over a single database hit the maximum capabilities of one instance
- There is a need of tenant physical isolation
Key features available in this moment
- Shard Map Management – It is used to create new shards, manage configuration, specify the key range for shards
- Data Dependent Routing – Enable to create connection to a specific shard and route that given request
- Shard Elasticity – Enables us to scale vertical (by control the type of SQL Azure that we are using) or horizontally (adding more shards/removing shards)
- Multi-Shard Queries – When a SQL hit multiple shards, the system will be smart enough to execute the request on multiple shards and merge the result for us.
- Split-Merge Service – When we increase, decrease the number of shards, the system will automatically balance the sharding data distribution
How to get a db connection
To get a connection to a specific database you can specify the ID of the entity. In this way you will get a connection directly to the database where your entity is stored.
ShardMapManager shardManager = ShardMapManagerFactory.GetSqlShardMapManager(
connectionString,
ShardMapManagerLoadPolicy.Lazy);
RangeShardMap<int> fooShardMapping = shardManager.GetRangeShardMap<int>("fooMapper");
SqlConnection fooDbConnection = fooShardMapping.OpenConnectionForKey(fooId, Configuration.GetCredentialsConnectionString(), ConnectionOptions.Validate)
Conclusion
This is a great feature, that can save your life. Now, sharding can be configured very easily by anyone.
This comment has been removed by a blog administrator.
ReplyDelete