Every app and data sometimes get corrupted, serious disasters are also possible (thankfully, much more rarely). While you read our blog, you probably do backups to protect the business in such cases. But since backup scenarios are different, there are significant differences in restoration. Let’s discuss possible ways of using cloud systems and summarize their pros and cons in order to find a truly wise data recovery.
Recovery Issues and Solutions
First of all we need to explain what is meant by words “recovery” and “disaster”. To get more detailed view on these concepts, feel free to check out our article about backup and disaster recovery difference. For now, let us just say that every restoration task can be split into these scenarios:
- Regular data restore (for example, user has accidentally deleted his work file)
- Urgent data restoration (the same as above, but it is a crucial data for business and it is important to get it back quickly)
- Disaster recovery (something hit your data center or your system and you have to restore the availability ASAP)
If we are talking about cloud solutions, there are three ways to handle the situation:
- Local to Cloud restore. Thus you have a backup in your local storage that could be used to restore the data to a cloud. Or you have a failed local server with local backup available - so it is possible to restore the server as a cloud VM.
- Cloud to Local restore. In this scenario you will download the data from a cloud backup to your local server. Often used while it is needed to restore archived data.
- Cloud to Cloud restore. Let’s call it “cloud native” restore - it is ability to restore a cloud system or data (VM or storage bucket\folder\file) using cloud-stored backup.
Look at these options closely.
Local to Cloud Restore
Local to cloud restore is used when you run some business system in the cloud and store its backups locally (for costs saving or as an additional protection level). This restore scenario has one main bottleneck - bandwidth of the Internet connection between your office and provider’s cloud (for example, Amazon S3\EC2).
It is suitable only for restoration of:
- Quite small user files in remote bucket or VM
- Non-crucial data that can wait for a few hours or days while transferring
- Initial data seeding. For example, you may want to setup a new cloud VM using old server’s local backup
The main advantage of this option is cutting of backup storage costs. It is not suitable for serious disaster recovery scenarios due to low Internet bandwidth to fast transfer of, for example, 200 GB of average server image. Let’s check the time needed to transfer this data to Amazon EC2 VM using common 100 Mbps connection:
- 200GB * 1024 * 8 = 1 638 400Mb
- 1 638 400Mb / 100Mbps / 60(for minutes) / 60(for hours) = 4,5 Hours RTO.
Cloud to Local Restore
It is a typical case of using cloud backups. You regularly pay for a backup cloud storage and use it from time to time to get some user file back. As a main benefit the Local to Cloud Restore lacks, you will not have to worry about backup storage durability and availability since it is all backed by cloud provider. Unlimited storage is also considered a great plus.
There are a few cons, of course:
- The same bottleneck as of local to cloud scenario — limited Internet connection speed.
- In case of disaster you need a working server hardware to restore the data on.
- You need to pay for cloud backup storage. But it is possible to cut the costs by using one of the “discount” storage tiers — for more info, check this article.
Cloud to local restoration is typically used as a long-term backup archive, but it can be handy to use it for small restores.
Note: Amazon has a great feature to relatively quickly download\upload large data to\from its cloud system — Amazon Snowball. To learn more, check our recent post.
Cloud to Cloud Restore
The best RPO\RTO solution for a customer with cloud-based main infrastructure. If you run a production VMs as, for example, Amazon EC2 instances, you can restore them very fast using cloud to cloud restore. It is possible because the backup data is stored in the same cloud system thus eliminate long transfers. The second case of using is to have a copy of bucket data, so it becomes possible to restore crucial user data within a few minutes.
Let’s summarize pros and cons of cloud to cloud restore.
- Fast recovery of a server images or user files.
- There is no need to have a server infrastructure in the company’s office.
- You don't have to worry about backup storage size, durability and availability.
- Quite expensive solution since you have to pay for VM (Amazon EC2) and storage (Amazon S3) cloud services.
- Higher data access latency (compared to local app installation) limits supported apps list. You may not be able to deploy services that are sensitive to the access latency or bandwidth.
This scenario can be used as a part of disaster recovery because you have geo-independent cloud infrastructure which is replicated over a few data centers across the country. Cloud to Cloud restore can also be used as an entry-level disaster recovery solution like in this case, when you have a standby cloud VM (it eliminates unnecessary costs) and a separate local service. To handle any local issues or disasters it is possible to simply enable standby VM and redirect user traffic to the cloud.
What Is the Truly Wise Data Recovery?
As it happens, there is no “silver bullet” for any data protection case. But you have a few modern backup options to handle with all data-related issues:
|Local to Cloud||Cloud to Local||Cloud to Cloud|
|Suitable for Disaster Recovery||Partially — too large RTO||No||Yes|
|Ability to Restore User Files||Yes||Yes||Yes|
|Initial Data Seeding||Yes||No||No|
|Local Server Infrastructure Required||No (if run a VM instance in the cloud)||Yes||No|
|Need to Support Local Backup Storage||Yes||No||No|
|Suitable for Latency-Demanding Apps||No||Yes||Partially - depends on traffic type|
We also suggest you checking the new CloudBerry Backup 5.8 so you could be able to restore an image-level backups directly to Amazon EC2 or Microsoft Azure VM (support of Google CE is coming soon). You can try CloudBerry Backup 5.8 absolutely for free during a 15-day fully functional trial period!
|Improve Disaster Recovery With CloudBerry Backup
||Try for FREE|