To aid in better understanding what Amazon Elastic Block Store is and, more importantly, what it is not, the following is an attempt to debunk what seems to be some common misconceptions -- at present time -- regarding the Amazon Elastic Block Store (EBS) solution.

Myth #1: Amazon Elastic Block Store was designed to solve data corruption issues.

Debunk: Amazon Elastic Block Store is persistent & redundant block level storage on demand. Nothing more, nothing less. If you're having issues with data corruption, Amazon Elastic Block Store isn't going to solve them for you. That's not what it was designed for.

Myth #2: Amazon Elastic Block Store intends to be the holy grail of SAN storage

Debunk: Amazon Elastic Block Store is not a SAN storage solution. It's persistent & redundant block level storage on demand. Nothing more. Nothing less. If you want to piece together a SAN on EC2 you'll need to do so using SAN tools.

Myth #3: Amazon Elastic Block Store cures cancer.

Debunk: Amazon Elastic Block Store is persistent & redundant block level storage on demand. Nothing more. Nothing less. It's not chemotherapy, nor DMSO, nor anything coming from the general direction of the Fred Hutchinson Cancer Research Center

Myth #4: Amazon Elastic Block Store is a distributed filesystem solution.

Debunk: Amazon Elastic Block Store is persistent & redundant block level storage on demand. Nothing more. Nothing less. If you want to piece together a distributed file system you'll need to do so using distributed file system tools such as LVM, NFS, AFS, OCSF2, etc.

Myth #5: Amazon Elastic Block Store cures headaches caused by LVM or distributed filesystems.

Debunk: Amazon Elastic Block Store is persistent & redundant block level storage on demand. Nothing more. Nothing less. If you have a headache, take aspirin, ibuprofen, Aleve, and/or other medications designed to help alleviate the suffering induced by a headache. DISCLAIMER: Just because my initials are MD doesn't mean I'm a doctor. Consult yours before proceeding with any type of treatment for LVM and distributed filesystems induced headaches.

Myth #5: With Amazon Elastic Block Store you no longer need to perform database backups.

Debunk: Amazon Elastic Block Store is persistent & redundant block level storage on demand. Nothing more. Nothing less. While Amazon Elastic Block Store block devices are significantly more durable than the local ephemeral disks and therefore the chances of losing data slim, you should *ALWAYS* make consistent and proper backups of missional critical data and store it on a medium *other* than the medium the database resides on. One important feature of Amazon Elastic Block Store that aids in this process is the snapshot capability which allows you to backup critical data to S3 with ease. After the initial creation and mounting of an Amazon Elastic Block Store snapshot the only thing you need to do to deepfreeze your data is make a call to the Amazon Elastic Block Store snapshot API which creates a diff snapshot (think CoW for those of you who understand what Copy on Write is all about) layering it on top of each preexisting snapshot, each of which are stored on S3. Restoring from any particular snapshot in time and/or building a new Amazon Elastic Block Store snapshot from a pre-existing snapshot (or groups of layered snapshots) requires nothing more than a call to the API and *POOF* you're up, running, and back in the game.

Myth #6: Amazon Elastic Block Store doesn't give you anything you don't get with S3.

Debunk: Amazon Elastic Block Store is persistent & redundant block level storage on demand. Nothing more. Nothing less. S3 provides persistence & redundance, it is not designed to be a mountable block level device. There are 3rd party solutions that provide this functionality, but each of them are required to utilize the HTTP/HTTPS-based S3 API for read/write operations which, quite obviously, comes with the overhead of HTTP/HTTPS. Amazon Elastic Block Store does not require you utilize an HTTP/HTTPS interface for read/write operations. You simply mount the device, if necessary format it with the file system of your choice, and then use it just as you would any other mounted device on your system.

That doesn't mean that alternative solutions will become obsolete with the introduction of Amazon Elastic Block Store. Each of these solutions have their advantages that go beyond that in which you gain with a traditional block level device. So with the addition of S3's persistence and redundance, they each have benefits that you don't gain with Amazon Elastic Block Store. But, like Amazon Elastic Block Store, these solutions come with a cost. Determining which service meets your specific needs in the most cost effective manner is an excercise I will leave for the reader.