The Cloud Paradigm Shift
Many companies try to utilize the cloud using the same techniques they used when hosting their own applications, or hosting it in a data center. Servers, IP Addresses, Machine Names and many other things are essential for success and maintenance.
In contrast, when using the cloud you tend to think in terms of functionality. Web services, applications services, data storage services, communication services, workflow. Instead of focusing on where your code resides, you instead focus on how to write your applications to work with a service hosted on one or more machines working as a single unit to the consumer.
Stephen Wynkoop provides a lot of great insight into this paradigm shift in his SSWUGtv episode, “The Great Amazon Power Outage 2012 – or, how to recover if you were hurt by the AWS outage, with Stephen Wynkoop”. You can find it by searching for the How To programs on our site.
I’ve been working with application fabrics and distributed database applications for nearly a decade now and come to appreciate how nice it is to manage software at a service level rather than a specific server. Most fabric like applications allow you to introduce and remove servers from the fabric without loss. Distribution of source code is handled by the fabric rather than deploying individually to each server. In short, you really don’t have to know what server is doing any particular unit of work.
Microsoft takes the position that since failover is virtually guaranteed, there is no need for backups, at least in SQL Server Azure. You really don’t know what server is hosting your database at any given point; and that is the best part. It may cause you to use different techniques when writing your code. But that is a small price to pay to gain the reliability of an elastic, extremely high fault tolerant system.
You can’t build your own fault tolerance at the same level for the nearly the cost of a hosted cloud system. The difference is, as Stephen says in his SSWUGtv editorial, if you use the Cloud like a hosted provider, you’re missing half of the power of going to the Cloud.
Cheers,
Ben
$$SWYNK$$
Featured Article(s)
Learning SSAS-Part IV: Deploying and Processing a cube
In this article I talk about the steps you need to take to deploy your SSAS cube and using it thereafter to run your queries. I also talk about the background processes that Analysis Server uses while you deploy and use your cubes.
Featured White Paper(s)
Ten Tips for Optimizing SQL Server Performance
read more)
Featured Script
dba3_ddl_CalendarSchema_Article
Modeling Date Logic III: Implementing Business Calendars (by example in MS Sql Server 2000) http://bitonthewire.wpengine.com/see/17379 Se… (read more)