First steps on EBS
A deeper view
Bearing in mind that EC2 local storage is not persistent, and that in case of hardware failure, you will lose all data stored in your instance, understanding AWS's persistent disk service, the Elastic Block Store is especially important. This short course will introduce you to provisioning EBS volumes, and using them for enhanced reliability as EC2-attached data and backup snapshots.
Note: some elements of this course are a bit dated. You might like to view the "Storage" video from the AWS Storage Fundamentals - AWS 140 course for an updated version.
And what happens if the maximum number of IOPS AWS can offer is not enough for you or if you need to store more data then the maximum volume size? Don't worry you can set up a Raid0 array in your Linux machine aggregating two or more EBS volumes. By putting multiple volumes in a Raid0 array, you will get a storage capacity that is the sum of all single volumes capacity and a throughput that is almost the sum of all throughputs. Look at this Instance we've attached here the three volumes we created awhile ago. Let's format the first volume and then mount it.
Now let's run "hdparm" and "ioping" to bench it's performance. Okay as you can see, it's providing around 150 IOPS and this is exactly what we would have expected. Now let's set up a Raid0 array with the other two drives and launch on them the test we just ran on the single volume. Can you see that? Volumes performance has just doubled.
In the last years Giorgio spent much of his time on the design and deployment of platforms based on new technologies like NoSQL Databases (MongoDB, HBase, Cassandra), Software Defined Networking (OpenFlow, Open vSwitch), Virtualization (Xen, Linux KVM, VMware) and Cloud Computing (OpenStack, Eucalyptus). He has a very deep knowledge of Amazon Web Services (AWS), RackSpace Cloud and related services, and has worked with many other IaaS / PaaS platforms (Google Cloud, Heroku, AppFog etc.).