Managing Allocation Requests
What is an Allocation Request?
The main organisational structure on the Melbourne Research Cloud is the
Project. When you create an instance, volume or object store container they
belong to a specific Project. Once you have logged into
the dashboard, you can change Project
Projects dropdown in the top-right navigation bar. The selected
Project will be what dictates the resource limits and in which subsequent
actions are made.
You may submit an
Allocation Request to apply for a new Project and request
an allocation (quota) of cloud resources. The Allocation Request may also be
amended at any time to request more resources, update any details or
extend the Project. Resources are allocated for a limited length of time to
ensure that the Melbourne Research Cloud is being actively used.
Any user may have, or belong to, any number of Projects.
Where do I create or manage my Allocation Requests?
Allocation Requests are managed in the Melbourne Research Cloud dashboard using the "Allocations" tab on the left-hand navigation panel:
- A new allocation request can be submitted at Allocations > New Request.
- Your list of allocation requests can be found under Allocations > My Requests.
Note: The only user who can see projects under My Requests is the user who created the allocation request. In other words, if a Project needs to be extended or updated in any way, that request needs to come from the Project owner.
Allocation Request Lifecycle
You can submit a new allocation request at Allocations > New Request.
See the below section Requestable Cloud Resources for help in completing the Cloud Resources section.
Allocation requests can take up to 2 weeks to process.
Your allocation request may be declined. If this occurs you will receive an email with advice regarding the reasons why. You may be asked to amend and resubmit your original allocation request.
If your allocation request is approved you will receive an email telling you that your project has been provisioned.
The start date will be updated to the date of approval if the approval occurs later than your specified start date.
Your project will have an end date. You will be reminded that your project will be expiring two weeks before the end date. You may extend the allocation request at any time in the dashboard.
If you do not submit an extension request:
- On the expiry date the allocation quotas will be set to zero. Existing virtual machines and storage resources will continue to be accessible, but no new resources can be created.
- Two weeks after expiry, all instances and data in this project allocation will be archived and safely stored.
- Three months after expiry, all archived instances and data will be deleted.
Requestable Cloud Resources
The following flavours can be launched on the Melbourne Research Cloud. Use this table to determine the number of Instances and vCPUs to include in your Allocation Request form. All instances come with a 30 GB disk.
|Flavour name||vCPUs||RAM (GB)||Suggested use|
|uom.general.1c4g||1||4||Simple web hosting|
|uom.general.2c8g||2||8||Database driven website|
|uom.general.4c16g||4||16||Data Science using RStudio or JupyterHub|
|uom.general.8c32g||8||32||Data science on larger data sets|
Note: the General flavors run on oversubscribed hypervisors.
If you require servers with higher memory or vCPUs than the general options above, there is an option to select Premium Flavours. You will be asked to justify your request. The Premium Flavours are
|Flavour name||vCPUs||RAM (GB)||Notes|
|uom.vdi.m10-6c30g||6||30||Perfect for Virtual Desktops|
|uom.gpgpu.k80-6c55g||6||55||General Purpose GPU|
|uom.gpgpu.2gpu-k80-12c110g||12||110||2 General Purpose GPUs|
Note: the Premium flavors run on dedicated hypervisors that are not oversubscribed.
A persistent volume looks and acts like a hard drive that can be attached to your virtual machine instances. Volumes and their data persist independently of virtual machine instances and volumes can be dynamically attached and detached to/from different virtual machine instances.
Object Storage is a large accessible online storage location that you can reach from most machines with internet connectivity. Object Storage requires cloud native applications for access.
The database service provides a simple interface for provisioning and managing database engines supporting MySQL and PostgreSQL, with MongoDB coming soon. You can specify the number of database servers, the total database storage your project would require, or both.
Virtual resources for building more advanced network topologies for your instances, including routers, load balancers and private networks.