I made time this week to attend the so-called ‘AWSome Day’ – a one-day seminar that Amazon has created in order to build familiarity with their ‘Amazon Web Services’ (or AWS) offering. While I’m no stranger to AWS (I was playing around with it a couple years ago, & have recently been working through the Wittig brothers’ Amazon Web Services in Action book), I thought that this would probably be a good way to pick up some worthwhile tips & ‘best practices’. I also hoped to gain additional familiarity with the ways that different components of the AWS universe can be combined to enable new services & new ways to address problems. (The fact that it was free was just a bonus.)
Our host for the day was Stephen Cole, a member of Amazon’s Technical Training group. While he certainly appeared to have a very solid understanding of the material, his presentation was marred slightly by sub-optimal acoustics – there were times when the speakers sounded somewhat “muddy” & unclear. It helped that the projected slides were quite clear & covered the topics in reasonable detail. (Actually, given some of the bad jokes that Mr. Cole delivered, perhaps muddy acoustics weren’t the worst thing that could’ve happened… heh.) After covering some basic Amazon & AWS history*, the demonstration cut right to the chase by giving us two hours that explained ‘AWS Infrastructure’ (with a generous break between the two hours).
Of course even with two hours, it wasn’t really possible to get into any real detail, since there are currently over 2,500 distinct services within the AWS “universe”. (I particularly appreciated the slides that showed how “traditional” datacenter-style computing infrastructure compares to the AWS cloud model.) Naturally ‘Storage Services’ were a part of the discussion, which included this tidbit – “Why use EBS over S3? When a small change to a file is made, EBS will change one block, whereas S3 will rewrite the entire file.” (Naturally there are use cases for which S3 is the appropriate choice, so YMMV…) After covering infrastructure, it was time for a catered lunch courtesy of Amazon & Intel.
Following lunch was the topic that I’d looked forward to – ‘Security, Identity & Access Management.’ One of the interesting things that Mr. Cole pointed out was that Security Groups are applied to users, roles, & instances (i.e., you can use them to configure firewall rules for instances) rather than ACLs, which apply to subnets. That would seem to allow for a good deal of flexibility. Another tidbit – the IAM component “sits next to” the hypervisor, which means that it can regulate AWS resources – but has NO visibility into operating systems, so it’s not the appropriate tool to use for anything within the OS.
The topic of ‘Databases’ under AWS was next up, & while DBs generally aren’t topics that generate much interest in me, the presentation must have been fairly engaging because I managed to learn a thing or two about SQL vs. NoSQL, & how each grows most efficiently.
After DBs, there was another generous break, which was followed by the day’s final topic, ‘Elasticity & Management Tools.’ I was mildly surprised at how interesting this topic turned out to be, the flexibility that can be gained by combining the monitoring of CloudWatch with Auto-Scaling & Elastic Load Balancing is pretty compelling for non-static workloads. I kind of enjoy playing with monitoring systems (i.e. Nagios or Zenoss), so I’m thinking I’d like to get my fingers into a CloudWatch instance, just so that I can see what it can do (especially related to the custom, user-created alerts…).
Finally, a small confession – While yes, I did get started working with AWS using the book mentioned above, I’ve recently put that on the back burner so that I could focus on learning the Python programming language. But the truth is that this “AWSome Day” has kindled (no pun intended!) a bit of enthusiasm in me, & I’m looking forward to digging back into it as soon as I complete the Python course that I’m working through. If an “AWSome Day” is coming to your town, & AWS isn’t something that you already use regularly, I’d recommend that you take a day to check it out. You might just pick up some worthwhile info!
* Did you know that ‘Amazon’ was initially called ‘Cadabra’ (as in “Abracadabra”)?? Apparently a lawyer thought that that name was too similar to the word ‘cadaver’, & suggested that they rethink the name…
