Virtual Datacenter

From Cloud Avenue
Jump to navigation Jump to search

Overview

A Virtual Datacenter (vDC) is a pool of resources including:

  • computing capacity, expressed in GHz or in vCPU, and in amount of RAM,
  • storage capacity for virtual disks of VMs,
  • one (or more) vDC Edge Gateway (T1), providing connections to external networks (internet or BVPN), as well as internal networks to the vDC.

Computing capacity (CPU + RAM) and storage are available according to several classes of services.


Notepad.png
Note!
A vDC corresponds to a service class and a resource allocation mode (PAYG or Allocation Pool). It is not possible to change the resource allocation mode of a vDC after its creation. If the Customer wishes to change the resource allocation mode of his vDC, then he must order a new vDC and migrate his vApps/VMs to it.


vDC specifications and performance classes

Service class Eco Standard High Performance VOIP
VM limits (vCPU / RAM / Storage) 4 vCPU / 16G / 4 TB 8 vCPU / 64G / 6 TB 40 vCPU / 128G / 6 TB 32 vCPU / 128G / 6 TB
Usage Windows VM

Dev, test, labs

All Big Data

Real time

Intensive computing

IPBX

Real time

Allocation mode PAYG & Allocation Pool PAYG & Allocation Pool Allocation Pool Reservation Pool
One Room
Dual Room
HA Dual Room
Billing mode PAYG, Reserved et DRaaS PAYG, Reserved et DRaaS PAYG & Reserved Reserved
vCPU frequency PAYG ou DRaaS 2,2 GHz 2,2 GHz 2,2 GHz 3 GHz
Reserved 1,2 GHz mini 1,2 GHz mini 2,2 GHz 3 GHz
Notepad.png
Note!
In the VOIP class of service, the DRS settings in place disables vMotion (the automatic movement of a virtual machine from one physical server to another) and VMware HA (the automatic restart of a VM hosted on a physical server down on another physical server), in order to be compatible with the operating mode of the majority of IP telephony solutions. However, for maintenance purpose, the DRS Fully Automated (vMotion) may be activated during maintenance window to empty the ESXi which needs an upgrade.


Availability classes

On the Val de Reuil Datacenter campus, the NGP platform is deployed in two rooms, each being completely independent of the other (energy, cooling, networks), which makes it possible to offer several classes of availability for the VM of a vDC:

Availability class Number
of vDC
Configuration VM location Mecanism Customer can choose the VM location Impact on loss

of a room

SLA

(vDC availability)

One Room 1 1 vDC with a single storage policy Not predictable the VMs are randomly distributed over the 2 rooms according to the capacity rules. no All or part of the VMs are stopped 99,95%
Dual Room 1 1 vDC with two storage policies (one per room)

example:

  • SILVER_R1 and
  • SILVER_R2
Room 1
and
Room 2
the VMs are located in a room thanks to the storage policy chosen to create them. yes VMs localized

in the faulty room are stopped

99,95%
Dual Room 2 2 vDCs each with a storage policy per room, for example:
  • 1 vDC with GOLD_R1
  • 1 vDC with GOLD_R2
Room 1
and
Room 2
the VMs are located in a room thanks to the storage policy chosen to create them. yes VMs localized

in the faulty room are stopped

99,95%
HA Dual Room 1 1 vDC with a single Dual Room type storage policy (storage replicated between the 2 rooms) Room 1
and
Room 2
the VMs are randomly distributed over the 2 rooms according to the capacity rules, on extended clusters (compute + storage) yes, with anti-affinity rules Shutdown / restart VMs

of the faulty room

99,99%


The detailed description of Availability Classes is here: Availability classes

vDC resource management

The resource allocation of a vDC is chosen by the Customer. This allocation is defined when ordering the vDC, then modified as often as needed by the Customer.

A vDC's resource allocation is a "physical" limit that VMs cannot exceed in order to run.

A VM will be able to start if the following two conditions are met:

  • ∑ vCPUs of VMs already started + number of vCPUs of the VM to start ≤ total allocation in number of vCPUs of the vDC
  • ∑ of RAM of VMs already started + amount of RAM of the VM to be started ≤ total RAM allocation of the vDC

Allocation mode of compute

The allocation of a vDC's resources to the VMs it hosts can be done according to 3 modes:

  • PAYG
  • Allocation Pool
  • Reservation Pool

PAYG

In this allocation mode, the resources are expressed in number of vCPUs (= Limit in GHz / vCPU frequency) and in GB of RAM. In a vDC configured in PAYG mode, the resource allocation of a VM is done at the VM level. When the VM starts, its vCPU and RAM allocation is subtracted from the total amount allocated to the vDC. A VM can never consume more than its allocation. This is a limit.

The billing method adopted is then pay-per-use (PAYG).

Allocation Pool

In this allocation mode, the resources are expressed in GHz and GB of RAM. These resources are a maximum usage limit with a guaranteed minimum reservation. Resource allocation for all VMs is done globally when creating the VDC. Unlike PAYG, there is no limit per VM and its consumption can be adjusted according to the needs and the total use of the VMs of the VDC (without exceeding the total limit).

For resource allocation (see § 6.1.3 above), the frequency of the vCPU is defined by the Customer in reserved billing mode, when ordering the vDC, with a minimum of 1.2 GHz.

The Allocation Pool mode allows the vCPU of a VM initially configured to consume up to the maximum value provided by a physical process (example: 2.2 GHz). This burst effect will activate under certain conditions:

  • A VM under heavy CPU load occasionally requires additional power
  • There are still GHz available (not consumed) in the vDC.

This allocation mode is recommended for vDCs whose VMs are permanently on.

Reservation Pool

In this allocation mode, the resources are expressed in GHz and GB of RAM. But unlike the Allocation Pool mode, it is a firm reservation and therefore just as much a guaranteed minimum as well as a guaranteed maximum reservation. The allocation of reserved resources (therefore guaranteed minimum) of a VM is done this time individually VM by VM by the administrator of the organization throughout the life of the VDC. It is a fine and total control by the administrator of the organization of his resource.

The resources allocated to each VM are 100% guaranteed for the VMs, which requires proper sizing when defining the VDC so that all VMs intended to work can start.

This resource allocation principle is used for VOIP-type vDCs, which allows telephony or videoconferencing applications to operate with the maximum possible isolation from their environment.

Notes

  • One would think that the "Reservation pool" mode is equivalent to the "Allocation Pool" mode where one would have reserved 100% of the limit. This is not the case because only the "Reservation Pool" mode allows reservation at the VM level.
  • However, if you want to reserve as much as the limit without ever having the need to reserve at the VM level, you can take an "allocation pool" mode (this is not recommended for VOIP type applications).

Billing mode of compute

Two billing modes are available.

Billing mode PAYG (per per use) & DRaaS Reserved
Resource allocation PAYG Allocation Pool

Reservation Pool

Billing resources allocated to the started VM 100% of the vDC resources

Billing mode "PAYG" & "DRaaS"

PAYG & DRaaS Compute RAM
Resources billed (WU) vCPU GB
Quantities invoiced Number of vCPUs allocated to each started VM

x

number of minutes / day

Amount of RAM allocated to each started VM

x

number of minutes / day

The time of use is counted by the minute.

A vDC in DRaaS mode is used when setting up VM replication between the Customer's private infrastructure (On Premise) and the Cloud Avenue platform.

Billing mode "Reserved"

In this model, all the GHz and RAM resources of the vDC are billed at a flat rate. This principle allows the Customer to benefit from more attractive rates.

Reserved Compute RAM
Resources billed (WU) GHz GB
Quantities invoiced Quantity of GHz allocated to the vDC
x
number of days / month
Quantity of RAM allocated to the vDC
x
number of days / month

Example

A vDC of 34 GHz and 20 GB of RAM is subscribed in Reserved mode. The monthly billing for this vDC is calculated as follows:

  • 34 GHz x unit price x 30 days x 24h
  • 20 GB of RAM x unit price x 30 days x 24h

This model also allows the client to choose their level of over-allocation. For example, by choosing a 1.7 GHz vCPU, it will be possible to start 20 VMs of 1 vCPU and 1 GB of RAM. And the most CPU-intensive VMs will then be able to draw on the GHz that are not consumed by the least active VMs (Burst effect), but within the limit:

  • GHz available globally in the vDC (34 GHz)
  • physical CPU frequency: 2.2 GHz.


Return to Services catalogue