CHAPTER 2 ■ ACCESS CLOUD TABLE STORAGE 63 userTableService.Insert(user as User); } Listing 2-26. Recursively Loop the Embedded Entity Object to Insert a Relational Data Set to Cloud Storage Tables virtual public bool Insert(ICloudEntity entity) { bool success = false; ICloudEntity dependency = null; try { dataTableContext.AddObject( dataTableContext.TableName, entity); dataTableContext.SaveChanges(); dependency = entity.GetDependencyEntity(); while (null != dependency) { CloudTableServiceFactory cloudTableFactory = new CloudTableServiceFactory(); cloudTableFactory.FactoryCloudTableService(dependency) .Insert(dependency); dependency = dependency.GetDependencyEntity(); } success = true; } catch { } return success; } • To delete a set of relational data entities from cloud table storage, a parent entity object is responsible for passing the partition key that is the RowKey of the child record. The list of dependency entities is the one we built up when constructing each entity instance . The highlighted lines from Listing 2-27 show how to accomplish this task. Listing 2-27. Recursively Loop the Embedded Entity Object to Delete a Relational Data Set virtual public bool Delete(ICloudEntity entity) { bool success = false; foreach (ICloudEntity entityType in entity.DependencyType()) { ICloudEntity dependency = QueryDependencyEntity(entityType, (entity as TableStorageEntity).RowKey); CHAPTER 2 ■ ACCESS CLOUD TABLE STORAGE 64 if (null != dependency) { cloudTableFactory.FactoryCloudTableService(dependency).Delete(dependency); } } try { dataTableContext.AttachTo( dataTableContext.TableName, entity, "*"); dataTableContext.DeleteObject(entity); dataTableContext.SaveChanges(); success = true; } catch {} return success; } • The implementation for the Update() method is relatively simple as shown in Listing 2-28. It is just a combination call to the Delete() and Insert() methods based upon the strategy updating cloud storage data in buckets. Listing 2-28. Updating Relational Entities in Buckets virtual public bool Update(ICloudEntity entity) { bool success = false; try { if (Delete(entity)) { success = Insert(entity); } } catch { } return success; } • Elaborately designed, tested, and tuned PartitionKey and RowKey are very important tasks for cloud table data modeling. To learn more about this topic from the Microsoft documentation search the web for “Azure Choosing a partition key is important for an application to be able to scale well.” If you are interested in the data center traffic characteristics you can find the Microsoft official documentation at http://research.microsoft.com/en-us/people/mzh/wren09.pdf. CHAPTER 2 ■ ACCESS CLOUD TABLE STORAGE 65 Summary In this chapter we covered a lot of ground; we examined quite a few aspects of cloud table storage. We saw how to access data in the cloud and worked through quite a few hints and tips to use when working with cloud table storage data access. There are many useful tools available to you, such as Fiddler 2, LINQ, and REST. Storing your data in the cloud would be useless if you couldn't alter it or remove it, so we covered this aspect of data storage next. We saw how to update data in buckets (a useful technique with cloud storage) and covered some best practices. Finally, we looked at relational storage in cloud table storage, which is definitely an option for some projects. C H A P T E R 3 ■ ■ ■ 67 Working with Cloud Queue and Blob Storage Azure Queue and Blob storage services are two basic forms of storage offered by the Azure framework. Another basic cloud storage offered by the Azure framework is table storage, which we covered in the last two chapters. In this chapter we are going to focus on Queue and Blob. Azure Queue messages can be listened for via event subscription. This feature makes Azure Queue a good candidate for building an event-driven distributed system. In this chapter I’ll provide the basic know-how for using these two types of storage, a description of how to use both storage types to build an event-driven distributed system, and a tool for you to load a large stream to cloud Blob storage. The first exercise is about the basics of how to create and use the Azure Queue. The second exercise is an example of cloud Blob storage. In this exercise we will use Azure Queue as a trigger to create a blob record when the queue receives a message. The third exercise uses both Azure Queue and Blob storage services to build an event-driven distributed system. Before we see the exercises, however, let’s see what Azure Queue and Blob Storage services are. Azure Queue Azure Queue provides a simple and asynchronous work dispatch mechanism. This makes Azure Queue a great message delivery tool that can be used to connect different components of a cloud application into an integrated service system. The outstanding merits of Azure Queue are high availability, durability, and performance efficiency. Azure Queue guarantees message delivery and ensures that a message can be processed at least once. Azure Queue provides REST interfaces, which allow applications written in languages other than C# to access the queue at any time from anywhere across the Internet. This makes a cloud application or on-premises application very easily integrated, extendable, and scalable. Azure Queue can be used for both cloud applications and on-premises applications for the following two purposes: • Message communication bus • Component or functional module decoupling