Can I create a repository in the cloud?

Currently neither AppAssure nor Rapid Recovery support the use of cloud storage for a remote repository. AA/RR is a very high IO application that require low latency, local storage in order to function properly. Our Product team is considering cloud storage for a repository and last I heard they were in the concept phase as they decide on how best to make use of cloud storage.

Currently the only options we provide for use of cloud storage is 1) archive to cloud or 2) replication to a VM in a cloud service provider like Amazon or Azure. With the new ability in RR to attach to an archive and perform a restore directly from a cloud archive, it makes archiving a much more viable solution for offsite data storage.

I would like to add to the conversation that you can archive older recovery points directly to the cloud and that the cloud archive behaves like a read only repository (you can attach it directly to your local core, without having to download and import the archive).
If you have a hybrid cloud and desire to protect the machines that already are in the cloud, you can build a cloud core -- although this is a trivial case.
Backing up directly to the cloud is a difficult proposition logistically wise and as such, as Tim said, it is not supported yet.

Can you expand a little more on this? I am wanting to implement either archiving to cloud or replication to a VM. I am trying to contact both Amazon and Azure for pricing, but it's not easy to understand which product is needed for them to quote. You mention that attaching to an archive is a 'much more viable solution', which has me wondering is this is a cost decision. How can I get specifics on these options?

I was asking if backups could be done to the cloud. The answer I got was no. Although, as I understand it, already saved backups could be transferred (archived)to the cloud, using the Rapid Recovery software. They can also be restored from the cloud. We are moving to do that but haven't completed that yet.

thanks. it's very confusing to understand which cloud solution/product you need for each method. For replication, it appears to be a VM on windows, so it's truly another core in the cloud. This would be ideal in my situation, but maybe it's cost prohibitive. For archiving, maybe it's a cheaper more realistic route. Love to hear what you've learned. thank you.

It really depends on how much data you are trying to store in the cloud and what your change rate is per day. I have a 4 TB repository on an Azure VM and that costs me about $100 per month. There is a Rapid Recovery VM template that you can purchase which comes with everything you need to configure the cloud VM (minus the license, you just use the same one for your local core on it). The benefit of a core in the cloud is that you can set a retention policy and minimize the total amount of storage you use.

With archiving there is no way to set a retention policy. If you configure incremental archiving the core simply appends the latest data to the archive each time it runs. There is no way to delete data from the archive without deleting the entire archive. So you have to manually manage the archive data and to clean up that data you have to delete the entire archive set and archive a new set of data.

Wow Tim, thank you for that very specific and helpful information. Exactly what I needed to point me on the right path. It sounds like a core on a VM is the way to go. The size of your repository is right at the same size as mine, so your cost number helps me out tons to even know if this is in the budget, which it is. So, do you have a local core that runs the backups, and replicate to the Azure VM after hours? Also, what is the exact Azure product/service that I need? And lastly, where can I purchase the VM template? Thanks again!

Oh, and did you do a price comparison between Azure and Amazon? Wondering what Amazon has to offer..

So I'm a support guy for Quest. I have a couple of cores in a lab environment that I use to replicate data to an Azure target. My change rate is fairly low at only about 20 GB per day. My network link is only 10 Mbps upload from the lab to the WAN so the maximum I can push up to Azure is pretty small too. I don't have a replication schedule so it pushes data up after each backup. I do have virtual standby set up on the Azure core to export data to Azure standby VMs.

To purchase the VM template you need to create an Azure Portal account (portal.azure.com) and then you'll be able to choose which Azure services you want to use. Here is our document on setting up the Azure replication target - support.quest.com/.../downloads. You'll want to search the Azure Marketplace for "Rapid Recovery" and you'll be able to purchase the Template VM there. As part of the purchase you have to choose a size for the VM which tells you the maximum resources available for that VM and the cost per month. My VM is a "Standard DS2 v2 (2 cores, 7 GB memory)". It's a little undersized for our minimum requirements, but it works. Just be aware that depending on your load you may need to increase your VM in size. That's pretty easy to do but you'll be limited in what sizes are available by the location you chose for the VM (all of that is documented by Azure). Make sure to read the limitations of the VM sizes as they can significantly impact you. Things like total network throughput, disk throughput, etc. I can say that I honestly haven't had any problems, but only have about 1 TB of my 4 TB repo used.

Once you've purchased the VM you'll be able to RDP to it. Before doing that though, I recommend adding disks to the VM so that you have the storage allocated and setting the public IP address to be static (otherwise it will change on you regularly). From there you can follow that setup guide I posted above to get the core functional and the storage added into a repository.

I have only priced an equivalent VM once in Amazon EC2. The number of options in EC2 are far more complicated, but far more flexible. I was looking at an m4.large instance and then 4 TB of elastic block storage (EBS). That would give you a VM with 4 TB of disk space. The cost is going to depend on how you pay for it. If you pay on-demand (meaning you are charged as it is in use) then it's more expensive than if you purchase for a full year. I priced the m4.large instance for a full year and it came out to about $115 a month. Adding on the EBS was an additional $185 a month. So total cost a month was about $300. Based on that, Azure looks to be the cheaper option.

***Disclaimer - this is only my personal research as I look through the options. I'm sure if you contacted Azure or Amazon with the specs you need, they may be able to get a better price. Both providers can change their pricing at any time, so it's best to look at the pricing prior to purchase and do your own comparison.