Nuno Filipe Godinho

PDC09 – Day Three – Building Scalable and Reliable Applications with Windows Azure

Brad Calder – Director/Architect

Data Building Blocks

• Types of Storage in Windows Azure
• Volatile Storage
• Local Storage
• Caches (eg. AppFabricCache and MemCache)
• Persistent Storage
• Windows Azure Storage
• Blobs
• Provide interface for storing name files
• Tables
• structured storage.
• Queues
• reliable storage and delivery of messages
• SQL Azure
• Storage Account Performance at commercial Availability
• Capacity
• 100TB
• Throughput
• Up to a few hundred megabytes
• Partitioning of Data Objects
• Load balancing is an internal concept of Windows Azure Storage
• Allows the system to automatically scale out access to your data to meet its peek traffic demands
• What matters to the application is the partitioning key used for objects
• All objects with the same partition key are always grouped
• Per Object/Partitioning Performance for Commercial Availability
• Throughput
• Single Queue and single table partition
• Up to 500 transactions per second
• Single Blob
• small reads/writes up to 30 MB/s
• large reads/writes up to 60 MB/s

Scaling Computation

• Compute Service Model – What is describes?
• The topology of your service
• Types of roles and their binaries
• How the roles are connected
• Configuration of the service
• How many instances of each role type
• Application specific configuration settings
• How many update domains you need
• Best Practices
• Due to application failures, upgrades or hardware failure
• Use multiple instances
• Queue Workflow Concepts
• Windows Azure Queue Provides
• Guarantee delivery
1. Worker dequeues Message and marks it as Invisible
2. Worker deletes Message when finished processing
• Note: If worker role crashes, message becomes visible for another worker to process
• A message may be processed more than once
• Assume messages put into same queue can be processed in any order
• Best Practice
• Make work items idempotent
• Scaling Queue Throughput
• Batch Work Items into Blob
• Group together many work items into a blob
• Queue up pointer to the Blob for the message
• Use Multiple Queues
• Job Manager
• Responsible form managing the execution of the Queues
• Work Items
• Continuation for Long Running Work Items
• Want to continue on failover
• High level approach
• Bread work item into smaller and repeatable steps
• Record progress

Lifecycle Management

• In-place Rolling updates
• Specify the number of upgrade domains in service model
• Breaks your roles evenly over the number of domains
• Rolling upgrade
• Walk each upgrade domain one at a time
• Upgrade just the roles in the current domain
• Benefits
• Minimizes availability loss
• Only one domain of roles are restarted at a time
• Allows local state to be uses across upgrades
• Caches application upgrade issues early
• Detect upgrade issues after first few domains
• Versioning with Rolling Updates
• Always assume you will have old and new running side by side in your service
• Version Everything
• Protocols, Schemas, Messages and everything else
• Windows Azure Tables Schema change
• Have a version property in each entity
• Types of Schema changes
• Adding a non-key property
• Perform similar step update process
• Update “IgnoreMissingProperties”
• Removing a non-key property
• Perform similar upgrade process
• Update “IgnoreMissingProperties”
• Changes in Partition Key or Row Key
• Most uses schema change is adding a non-key property