CumulatedCapacity
Definition
Signature
cumulatedCapacity(s:set<server>, r:string, nb:number)
s: a non-empty set of servers for a meaningful constraint. Servers not in theOnlinestate are ignored.r: a resource identifier such asvm,mem,ucpu,pcpuornodesto identify the number of virtual machines, the physical memory, the computational capacity, the physical CPUs, respectively.nb: a positive amount of resources.
The cumulatedCapacity constraint restricts to a maximum of nb, the total amount
of a specific resource of type r that can be used on the online servers in s to run VMs.
Classification
- Primary users: datacenter administrator
- Manipulated elements: VM placement, Resource allocation
- Concerns: VM-to-server placement, Resource management
Usage
The cumulatedCapacity constraint enables first the datacenter administrator to control
the shared resources of
a platform. As an example, each VM has at least one IP address to be accessible from the network.
In practice, a datacenter has a finite pool of addresses to share among all the VMs.
Such a datacenter has then a global hosting capacity limited by the size of the address pool.
In this setting, one cumulatedCapacity constraint may be used to limit the hosting capacity
of VMs according to the size of the address pool.
The cumulatedCapacity constraint may also be used to control license restrictions.
As an example, on a datacenter running vSphere, the hosting capacity is limited by the cumulated amount of
memory allotted to the running VMs. [43] In this setting, one cumulatedCapacity
constraint may be use by the datacenter administrator to guarantee the overall consumption of memory used
on the servers running VMWare is necessarily lesser to the maximum allowed by the acquired licenses.
Example
Figure 17 depicts a sample reconfiguration between a source and a destination configuration where each server provides 8 unit of CPU and 7 unit of memory resources to VMs. Each VM is associated to a gray rectangle that denotes its resource usage. In this setting, the following cumulatedCapacity constraints were considered :
figure 17: A reconfiguration motivated by cumulatedCapacity constraints.
cumulatedCapacity({N1, N3},7,"mem"). This constraint was not satisfied in the source configuration as the cumulated memory usage onN3andN1was equals to 9. This violation was fixed by relocatingVM4toN2.cumulatedCapacity({N3, N4},3,"vm"). This constraint was not satisfied in the source configuration as there was 4 VMs running on the two nodes while the constraint restricts this number to 3 at maximum. This violation was fixed with the relocation ofVM4toN2.cumulatedCapacity({N2, N4},16,"ucpu"). This constraint was satisfied in the source configuration as the cumulated UCPU usage ofN2andN4was equals to 12. The constraint is still satisfied in the destination configuration as the cumulated UCPU usage equals 16.
See also
Related Constraints
singleCapacity: A constraint to restrict the resource capacity on each server. AcumulatedCapacityconstraint is then equivalent to asingleCapacityconstraint when the set of servers incumulatedCapacityis a singleton.
Specialization(s)
To
singleCapacity:∀n ∈ ns, cumulatedCapacity({n}, nb, r)↔singleCapacity(ns, nb, r)