No products in the cart.
One important thing that one should understand while discussing capacity planning, is the
type and extent of responsibility for a designated level of service provider, as well as for a service
consumer. The end users, who only use SaaS facility, are customers of the SaaS category service
providers. The SaaS category service provider may in turn be customer of some PaaS category
service provider, if not an independent computing vendor. Similarly, a PaaS category service
provider is either an independent vendor, or consumer of IaaS category service provider.
To explain the whole scenario in a simple form, let us consider three different companies
who are providers of different cloud services under some agreement. Company ‘A’ – IaaS facility
provider, company ‘B’ – PaaS facility provider, and company ‘C’ – SaaS provider (Figure 10.1).
The customers of company ‘C’ are application users (end users of computing). They can
take three ways to inform ‘C’ about capacity requirements. Case-1) can estimate their exact
business requirement and inform ‘C’. Case-2) Pass the entire responsibility of supporting
their demand to company ‘C’. Case-3) can take a middle path where business demand beyond
a certain (estimated) level will be supported by ‘C’. The choice and requirements have to
be mentioned as contract in SLA, so that ‘C’ can estimate capacity requirements and take
appropriate measures to arrange them. Company ‘C’, after accumulating all SLAs made with
its consumers has to estimate the future demands in a similar way. They also have the three
similar choices like its consumers.
Company ‘B’ has two options to maintain resource capacity to support their business
demand. With proper estimation, ‘B’ can keep few additional virtual servers (supplied by ‘A’)
under their hold as reserved resources, to support any unexpected raise in load to manage the
capacity themselves. The other option is to directly pass the ball (the responsibility of capacity
planning) to company ‘A’ keeping faith on their capability. They can take the middle path also,
by managing the capacity planning up-to a certain level by themselves and then passing the
responsibility to the infrastructure service provider in extreme cases. But, a reputed service
provider must handle the responsibility on their own, without depending on others.
The physical infrastructural resource management is the responsibility of the IaaS service
provider ‘A’. They have no option to pass the capacity planning task to other service providers
as providing physical resources is the ultimate responsibility of the IaaS provider ‘A’. The entire
scenario has been summarized in Figure 10.2.
Hence, the capacity planning task needs appropriate estimation of future demands at each
layer of cloud services. Although SaaS and PaaS consumers can not directly participate in capacity
planning activity, they should estimate business demand and inform their respective underlying
layers regarding future demand in advance. Each layer must remain well aware about possible
future demand to keep them-self ready.
If company ‘B’ remains uninformed (from its upper layers) and fails to support resource
demand beyond their planning, then it is likely that ‘A’ may fail to support the feature load
in many instances, if multiple among their consumers (like ‘B’) fail to do so and. IaaS facility
provider should be seen as the last line of defense, and the ball (the responsibility of supplying
infinite resources) should be not be allowed to pass on there.
One important thing that one should understand while discussing capacity planning, is the
type and extent of responsibility for a designated level of service provider, as well as for a service
consumer. The end users, who only use SaaS facility, are customers of the SaaS category service
providers. The SaaS category service provider may in turn be customer of some PaaS category
service provider, if not an independent computing vendor. Similarly, a PaaS category service
provider is either an independent vendor, or consumer of IaaS category service provider.
To explain the whole scenario in a simple form, let us consider three different companies
who are providers of different cloud services under some agreement. Company ‘A’ – IaaS facility
provider, company ‘B’ – PaaS facility provider, and company ‘C’ – SaaS provider (Figure 10.1).
The customers of company ‘C’ are application users (end users of computing). They can
take three ways to inform ‘C’ about capacity requirements. Case-1) can estimate their exact
business requirement and inform ‘C’. Case-2) Pass the entire responsibility of supporting
their demand to company ‘C’. Case-3) can take a middle path where business demand beyond
a certain (estimated) level will be supported by ‘C’. The choice and requirements have to
be mentioned as contract in SLA, so that ‘C’ can estimate capacity requirements and take
appropriate measures to arrange them. Company ‘C’, after accumulating all SLAs made with
its consumers has to estimate the future demands in a similar way. They also have the three
similar choices like its consumers.
Company ‘B’ has two options to maintain resource capacity to support their business
demand. With proper estimation, ‘B’ can keep few additional virtual servers (supplied by ‘A’)
under their hold as reserved resources, to support any unexpected raise in load to manage the
capacity themselves. The other option is to directly pass the ball (the responsibility of capacity
planning) to company ‘A’ keeping faith on their capability. They can take the middle path also,
by managing the capacity planning up-to a certain level by themselves and then passing the
responsibility to the infrastructure service provider in extreme cases. But, a reputed service
provider must handle the responsibility on their own, without depending on others.
The physical infrastructural resource management is the responsibility of the IaaS service
provider ‘A’. They have no option to pass the capacity planning task to other service providers
as providing physical resources is the ultimate responsibility of the IaaS provider ‘A’. The entire
scenario has been summarized in Figure 10.2.
Hence, the capacity planning task needs appropriate estimation of future demands at each
layer of cloud services. Although SaaS and PaaS consumers can not directly participate in capacity
planning activity, they should estimate business demand and inform their respective underlying
layers regarding future demand in advance. Each layer must remain well aware about possible
future demand to keep them-self ready.
If company ‘B’ remains uninformed (from its upper layers) and fails to support resource
demand beyond their planning, then it is likely that ‘A’ may fail to support the feature load
in many instances, if multiple among their consumers (like ‘B’) fail to do so and. IaaS facility
provider should be seen as the last line of defense, and the ball (the responsibility of supplying
infinite resources) should be not be allowed to pass on there.