Question

3) Dublin-based Cloudenci is considering a new public cloud offering that assures a maximum time to repair (MTTR) of 600 seconds for any given failed component in its cloud infrastructure. Assuming a maximum of one failure per year:

a) Calculate the annual availability without planned downtime

b) If Cloudenci wants to offer a five nines SLA based on the availability, what is an acceptable MTTR assuming 1 failure per year with no planned downtime?

c) Discuss one of the following design approaches and its impact on failure in clouds: Service degradation, Buffering, Automation

Answer 1

Hi,

Please find the answer below:

a)

Assuming a normal year has 365 days

Number of seconds in a year = 365*24*60*60

1 failure in a year means MTBF = 1 year

Given MTTR = 600 seconds

%

Availability = 99.998%

----------------------------------------------------------------

b)

Given

Availability = 99.999% or 0.99999

1 failure in a year means MTBF = 1 year

We have to find out MTTR.

Simplify the terms

seconds

------------------------------------------------------------------------

c)

Automation can be used to detect failures in the cloud. Using automation we can perform continuous testing
of a cloud service. Continuous testing allows to configure monitors or alerts on the cloud service to the stake holders. Automation allows detection of failures in cloud service without too much manual effort.

Hope this helps.