I would start with work related to Service Level Agreements (SLAs). In any cloud providers offer them to users as an unwritten guarantee of the availability of their resources. There is also a lot of work regarding High Availability which may be a good place to start. One of the most common techiques is the m+n machines approach however others based solely on software exist.

We did some work and a product implementation based on the ideas at a start-up company I worked for summarized here - http://www.assurancestorage.com/objects/whitepapers/AssuranceFDIR.pdf, which was patented - http://patft.uspto.gov/netacgi/nph-Parser?Sect1=PTO2&Sect2=HITOFF&p=1&u=%2Fnetahtml%2FPTO%2Fsearch-bool.html&r=1&f=G&l=50&co1=AND&d=PTXT&s1=8473779&OS=8473779&RS=8473779  Disk drives (spinning rather than solid state) are definitely one of the most like to fail components in the cloud and will continue to be used for a long time based on cost per gigabyte (compared to solid state solutions).

System resilience It can provides redundancy and operational resources reallocation of resources.

If it is not about safe equipments and control algorithms to correct that.

My conviction is that all kinds of reliability under fault conditions can only be realized through the introduction of feedback.

In such complex systems razvletlennyh neobhomo have current information on the status of processes.

I wrote an article, but why is it no one understands. Probably, because the introduction of feedback involves the use differenitsalnyh equations in dynamic control.

Furthermore, there is a danger load machining time.

Look, there is a list of references.

I can advise you some work on modeling the observed linear systems for cloud computing.

Perhaps you at first glance, it's not about that. But it is done in usitemah management. When one line fails, it switches over to another line. And for this, it must exist.

http://0b4af6cdc2f0c5998459-c0245c5c937c5dedcca3f1764ecc9b2f.r43.cf2.rackcdn.com/11644-feedback13-luna.pdf

Article Concept of Feedback in Future Computing Models to Cloud Systems

Conference Paper Optimal Control of Applications for Hybrid Cloud Services

A few references

http://controlofsystems.org/FeedbackComputing2012/papers/Sess13-feedbackcomputing2012.pdf

http://www.ida.liu.se/labs/eslab/publications/pap/db/serra_TECS14.pdf

http://weblidi.info.unlp.edu.ar/WorldComp2013-Mirror/p2013/GCA3178.pdf

http://www.activecomponents.org/bin/download/Documentation/Publications/cloud.pdf

http://www.hindawi.com/journals/mpe/2014/816518/

Resilience is considered as an enforcement model for security and privacy in business networking. Cloud computing is an example of the considered systems. We have published the special issue "Security and Privacy in Business Networking" of the Springer "A" ranked journal Electronic Markets. Please find the article on resilience for security and privacy in business networking and all articles of this special issue via the attached one

Article Security and Privacy in Business Networking