Google Groups

Re: [ Cloud Computing ] 5 Key Lessons for Customers of the Cloud


Miha Ahronovitz Apr 23, 2011 1:05 PM
Posted in group: Cloud Computing
Well said Ray.

Miha

On 4/23/2011 12:10 PM, Ray wrote:

The other key is to avoid using cloud provider specific features like RDS. This is killing some AWS customers as RDS seems to be having an especially hard time recovering and there is no RDS equivalent in other cloud provider feature inventories to fail over to.

 

From: cloud-c...@googlegroups.com [mailto:clou...@googlegroups.com] On Behalf Of Khazret Sapenov
Sent: Saturday, April 23, 2011 11:47 AM
To: cloud-c...@googlegroups.com
Subject: Re: [ Cloud Computing ] 5 Key Lessons for Customers of the Cloud

 

 

On Sat, Apr 23, 2011 at 1:09 PM, CloudSigma <rob...@cloudsigma.com> wrote:

Following the major outage that AWS suffered in their east coast US
facility this week, after the dust settles, what lessons can customers
actually learn from the events of this last week?

Here are the five key lessons we've highlighted to customers:

Lesson 1: Both Cloud and Dedicated Computing Have Single Points of
Failure

Lesson 2: Size is No Protection from Outages without Redundancy

Lesson 3: All Data Centres Are Not Equal

Lesson 4: The Price-Performance-Reliability Metric

Lesson 5: Achieving a highly robust set-up is cheaper and easier in
the Cloud

Customers need openness from vendors about their infrastructure
choices and locations in order to create price-performance-reliability
comparisons between clouds. This is a key development needed if people
are to make the right decisions and create the appropriate strategies
in line with their computing needs in the cloud.

 

 

Customers already have wide choice of locations within Amazon EC2.

If you saw their status of availability, only one location of many was affected,

so if one has to engineer his/her apps for redundancy/resilience, 

then first quick and easy way would be to implement such DR functionality within one cloud (be it AWS or another provider, if they have multiple locations at all),

where you have uniform interfaces and formats, so re-engineering efforts do not sacrifice reliability and interoperability of final solution.