Storage

This page also exists in French.

Overview

Cloud Avenue provides 2 types of storage:

• Datastore type storage, for virtual disks of VMs (vmdk files)

• Network storage, accessible via NFS or SMB

Datastore type storage is made available to VMs in a vDC through storage profiles, from which the VMs can draw to provision their disks. The size of a vmdk file is limited to 2 TB. When a VM must have storage greater than 2 TB, the vmdk files of the VM must be multiplied to reach the target size.

Datastore type storage is provided in two modes:

• Shared datastore (by default),
• Dedicated datastore, with guaranteed performance.

Shared Datastore

Service classes

For each vDC performance class, one or more storage performance classes can be associated; the compatibility matrix describes them in the following table:

Services classes	Maximum IOPS	Available for the vDC
		Eco	Standard	High Performance	VOIP
Silver	600 IOPS / TB	▲	▲	▲	▲
Gold	1000 IOPS / TB	▲	▲	▲	▲
Platinum 3K	3000 IOPS / TB	△	▲	▲	▲
Platinum 7K	7000 IOPS / TB	△	▲	▲	▲

Some classes of service are available in Dual Room to offer room choice, and in HA Dual Room on a storage Metro-cluster, spread over two rooms in Val de Rueil. The characteristics and availability of these classes of service are detailed in the following table:

Services classes	Available for the vDC
	One Room	Dual Room	HA Dual Room
Silver	▲	▲	△
Gold	▲	▲	▲
Platinum 3K	▲	▲	▲
Platinum 7K	▲	▲	▲

Multiple storage profiles can be associated with a single virtual machine.

The IOPS resources are pooled between all the VMs using the Datastore on which the Customer deploys his VMs.

Billing

PAYG

The storage considered is the maximum space occupied during a day by:

• each VM, including the space allocated to a VM (the disk size) and by the snapshot possibly made by the Customer
• the Templates and ISO images present in the Client's private catalog

VM technical files like vmware swap file also consume space but are not charged.

The storage sizing of a vDC ordered by the Customer must take into account:

• the virtual disks of the VMs that will be created in this vDC
• VM RAM
• possibly the space to take a snapshot.

Per subscription

The storage taken into account is the observed average storage allocated to the vDC during the reference month.

Dedicated Datastore

A storage policy, made up of one or more dedicated datastores, is made available to the Customer for their exclusive use. That is to say that only the VMs chosen by the Customer will be able to use this storage space, which allows him to make the best use of the number of IOPS available.

The maximum size of a Datastore is 10 TB, but a storage policy can have several Datastores, and can therefore reach very large sizes. Contact support for requirements above 30 TB.

A storage policy is attached to a single vDC.

Service classes

This type of storage is available, depending on the vDC classes:

• In One Room
• In Dual Room (choice of room possible)
• In HA Dual Room (on a storage Metrocluster on both rooms)

Ordering a dedicated storage policy (dedicated datatore) is possible for all of the following classes of service:

Service classes	Performance	Minimum size	Available for vDC
			One Room	Dual Room	HA Dual Room
Silver	600 IOPS / To	8 To	▲	▲	△
Gold	1000 IOPS / To	6 To	▲	▲	▲
Gold Metrocluster	1000 IOPS / To	6 To	△	△	▲
Platinum 3K	3000 IOPS / To	4 To	▲	▲	△
Platinum 3K Metrocluster	3000 IOPS / To	4 To	△	△	▲
Platinum 7K	7000 IOPS / To	4 To	▲	▲	△
Platinum 7K Metrocluster	7000 IOPS / To	4 To	△	△	▲

À noter !
For a dedicated Datastore, the deduplication mechanisms embedded in the storage equipment benefit the Customer. That is to say, if the Customer deploys several VMs based on the same template, this will significantly increase the rate of deduplication, and consequently will limit the total amount of storage billed to the Customer.

Billing

The storage taken into account is the observed average storage reserved by the Customer during the reference month.

The quantity of IOPS actually provided corresponds to the volume of storage ordered times the number of IOPS/GB of the class of service subscribed.

Example:

4 TB of Platinum 7K storage can permanently use 28,000 IOPS, regardless of the number of VMs using this storage.

Network storage

Shared NAS storage is available for virtual machines and is based on Netapp technology.

iSCSI is not yet available directly on the shared storage portal.

General rules for network storage

• Volumes can be either Unix or Windows (no mixed mode)
• Customers can create up to 6 volumes per SVM between 500 GB and 10 TB in self-care. For additional volumes or volumes greater than 10 TB, the request must be made via a ticket. Volumes larger than 10 TB will have the same QOS policy as 10 TB (capping).
• Sharing/exporting is done at tree level.
• A Windows share is created on a Windows qtree on a Windows volume.
• An export is created in a Unix qtree on a Unix volume.

Shared storage portal (France VDR only)

A dedicated shared storage portal is available to the customer.

This portal offers the following functionalities:

• NFS v3 sharing
• SMB/CIFS sharing with local user management for file access. Active Directory integration.
• Snapshot management. Customers can choose one of the following snapshot policies to schedule snapshots and retention on their volume
• Volume provisioning (limited to 6 x 10TB volumes) and capacity visualization.
• SVM auto-provisioning

APIs are REST API compatible.

Storage classes

3 different storage classes are available to meet all usecases:

Network storage backup

Network backup can be performed on one or two sites.

Backup mechanism is as follows:

• A replication VLAN is required to link all Netapp racks on each site. This VLAN will be propagated on dedicated ICL ports.
• The VDR and the CHA replication VLAN must be connected via L3 routing to enable replication between sites.
• A DR SVM is created on the backup bay.
• Replication between the 2 SVMs takes place via the intercluster link.
• SVM DR or volume-to-volume snapshot replication is implemented (TBD).
• The SVM DR has connectivity on the same VLAN as the shared storage SVM.
• Replication and retention are applied on the basis of a list of predefined backup policies.
• Restoration is performed by read-only mounting of the DR volume and copying via the VM to the original location.

On each site, the vlan can be different. As the SVM ports are named after the VLAN, it's easy to know which VLAN on site A is to communicate with site B. Routing is acceptable for a customer wishing to access a backup SVM on another site. This is backup data, not production data.

Return to Services catalogue