Cloud Patterns - Channel 9

Using Windows Azure to Build Pop-Up Labs in the Cloud: Cartoon

Planky — Sat, 13 Apr 2013 10:41:47 GMT

Although you might build say, a Sharepoint infrastructure in Windows Azure and leave it deployed as a permanent feature, totally integrated into your enterprise infrastructure, there are a growing number of organisations taking advantage of the temporary nature such infrastructures can have. Whether they need temporary access to a fully functioning IIS infrastructure, SQL infrastructure, AD infrastructure, System Center infrastructure or even a fully functioning, full-fidelity instance of one of their home-grown LOB enterprise apps - they can all be "popped up", used, then "popped down" again.

Whether you're an independent consultant who needs to replicate a problem at one of your clients, a small business that needs to run a training course for its employees or a large multi-national corporation that wants a dev/test environment that exactly mimics their production environment - pop-up labs can provide these services.

This video explains how in some cases, fairly extensive pop-up labs can be deployed for free and shows how they are so economically viable because they are only paid for when they are "popped up".

You'll see how to use Virtual Networks, Windows Azure Virtual Machines and Cloud Services to build your own pop-up labs.

Windows Azure Service Bus Application Patterns

Haishi Bai, Abhishek Lal — Tue, 05 Feb 2013 15:49:59 GMT

Service bus is a very capable service provided by Windows Azure. In this episode we'll explore some of the application scenarios of two major service bus offerings – relay and messaging. We'll examine different characteristics of the two services in terms of differences in payloads, topologies, scaling patterns, and many other aspects. The following table is taken from the table we built during the session. We hope this episode can provide you some guidance and ideas when designing your own applications.

	Relay	Queues/Topics
Architecture	Coupled	Decoupled
Main value	Connectivity/location separation	loosely-coupled integration
Communication pattern	request-response, pub/sub	asynchronous, batch, offline, pub/sub, request-response possible
Payload	Ephemeral entities	Durable messages
Latency	Latency <10ms	Latency ~200ms
TTL	TTL < 1min	TTL may be in years
Connection	Concurrent listeners with high rates of connect/disconnect	Listener with long-lived durable connections
Scaling	Scale out to multiple listeners and then multiple relay endpoints	Scale out with more entities and data sharding
Topology	Fixed topology	Variable topology
Applications	Real-time data (telemetry) Distributed systems Event feed Connectivity to legacy systems/databases Live-status monitoring Multi-player games (interactive – Pong)	Durable data (orders) SOA/Service integrations Chat Load-leveling State synchronization (distributed) workflows Multi-player games (turn-based – Poker)
Development	WCF/HTTP	.Net client/REST API

Using the Request-Acknowledge-Push Pattern to Display Progress of Long Running Tasks

Haishi Bai — Thu, 17 Jan 2013 19:52:28 GMT

Many web sites need to deal with long-running tasks. However long-running tasks don't
play very well with the HTTP request-response paradigm. In this episode we'll
go through a very simple pattern: Request-Acknowledge-Push that enables a
simple, efficient, and scalable way of dealing with long running tasks.

Source code of this episode can be found at: http://azure4fun.blob.core.windows.net/code/CloudPatterns1-RequestAcknowledgePush.zip

You'll need to update both Web Role configuration and Worker Role configuraiton to use your own Service Bus namespace.

And the blog article mentioned in the video is: http://haishibai.blogspot.com/2012/12/dealing-with-long-running-jobs.html