In part 1 of this post I discussed the use of availability zones and regions to have fault tolerance for an application such that if an AWS or Azure data center goes offline you can still operate your application.
But redundancy and fault tolerance work on many levels in AWS and Azure. If one of your virtual servers goes down and you have a good backup system and process in place you can quickly recreate the server. The cost to you is primarily downtime. However, if there is a problem with the data storage mechanism you could lose actual data and that may be a much bigger problem. So, let’s look at redundancy in data storage.
In on-premises servers hard drives have RAID levels, which means the data on a hard drive us duplicated across other hard drives so that if a hard drive fails you don’t lose data. There are 3 main levels of data storage redundancy in AWS and Azure. In Azure there is Locally Redundant Storage (LRS) which replicates your data 3 times within a single data center. The next level up of redundancy is Zone-Redundant Storage (ZRS) which replicates your data across 3 availability zones in a single region. Lastly there is Geo-Redundant Storage (GRS) which replicates your data to another region at least hundreds of miles away from the primary region. The cost and reliability (Fault tolerance) of these storage systems both go up as you move from LRS to ZRS to GRS.
AWS also has different types of storage with different levels of redundancy. EBS (Elastic Block Store) data storage is what you use to make virtual hard drives. EBS is a high-performance block storage service designed for use with the AWS EC2 (Elastic Compute Cloud) service, which is the service you use to create virtual machines. It stores data in blocks much like a typical server hard drive does. EBS volumes are replicated across multiple servers in a single Availability Zone to prevent the loss of data from the failure of any single component.
Amazon S3 (Simple Storage Service) is an Object Storage service. An object storage service doesn’t work exactly like a virtual hard drive like EBS storage does. It is lower cost and lower performance that EBS but also more redundant. S3 storage replicates data into multiple availability zones whereas an EBS volume lives entirely in one availability zone. Note that S3 has different storage classes with different performance characteristics so you can select the combination of price and performance that best suits your needs.
By combining the redundancy of Azure and AWS storage technology with a good snapshot strategy that ensures frequent and consistent snapshots you can achieve the level of data storage fault tolerance that you need.
A managed IT services company like GizmoFish who is experienced with Biotech and Life Sciences companies and is knowledgeable with cloud platforms like AWS and Azure can help you either evaluate cloud systems who use these platforms or build your own infrastructure on these platforms.