Cloud Management - Deploying and Managing a Cloud Infrastructure - page 120

Capacity Management Capacity management is about the management of capacity for
present and, especially, future demand. The aim is to strike a balance between scalabil-
ity and usage, that is, between having too much unused capacity and not being able to
scale with rapid spikes in demand. Capacity management relies on two activities, namely,
change monitoring and trending. Change monitoring, as the name suggests, involves
monitoring for changes that may affect the capacity of a system. For example, when hard-
ware like storage and computing resources is upgraded, there will be some downtime, not
of the whole system but some of its parts and resources, which will lessen the capacity of
the entire system during the duration of the upgrade process. Capacity management takes
this into consideration and ensures that capacity is available as expected through other
means. Trending, on the other hand, looks at historical data and monitors current events
to predict spikes or valleys in usage and ensure that there is always enough capacity, even
when there is an extreme load on the system.
Systems Life Cycle Management Life cycle management in cloud computing is tied to
the life cycle of services rather than tangible products. The ITIL framework specifies five
phases: service strategy or planning, service design, service transition, service operation,
and continual service improvement. The phases are executed in sequence, except for contin-
ual improvement, which is part of all phases because the cycle goes back to the beginning,
where lessons learned are applied to the next cycle.
Maintenance Windows To a cloud computing provider, scheduling maintenance (and
how long it will take) is important so that customers will not be burdened and their busi-
nesses interrupted. As a customer, you can actually set the dates for downtime in your SLA.
Depending on the number of nines (99%, 99.9%, 99.99%, etc.) you specify, you can tell
your provider which periods can be used as downtime for maintenance of the hardware
where your system happens to reside.