Playing the Cloud: Why I Use AWS
I like AWS because it’s like playing the world’s greatest modular synthesizer.
Modular synthesizers are composed of separate units or modules with different functions that impact on sound. These modules can be connected together with patch cords (basically wires), a matrix patching system, or switches by the user to create a patch. A ‘patch’, basically, is some kind of unique sound that the synth can make once the modules are wired together under a specific configuration.
The best thing about synths is that they can make a lot of different sounds, many of which cannot be made elsewhere. The multiplying combinatory effect of possibilities when configuring and wiring up different modules is what creates this vast multiplicity of sound.
The same is true of AWS. Except now, instead of producing sounds, we are looking at producing computational results, artifacts, web applications, or whatever... This is achieved by configuring, then connecting different modules. The combinatory effect of possibilities here transforms IT, just as the synth transformed popular music. It makes it possible to quickly produce new results, products, systems, ideas etc.… unlike any others previously seen. And there is a multiplicity of potential in these, making it AWS of the best tools for IT innovation that there is.
The Modules
There are a lot of AWS Services. They cover databases, AI, computation, messaging, APIs. Basically it’s all of IT in a box, at your fingertips, ready to be played, by one developer, one team or many. Because it is cloud, it’s easy to experiment. Start up some services, wire them together, and see what kind of product/artifact/computation (i.e. sound) results.
A Random Example
To give a random example, we need to start with the equivalent of a unique sound. Let’s say I have an investment document and I want to store it in a database, but I want an automated element which will do all the heavy lifting for me.
What might that system look like? Well, I could take some pictures of the document with my phone, then have those analyzed and the text extracted and stored automatically.
I can get that result by configuring and wiring up some AWS services:
I can run up a simple front end application to the cloud using the Amplify Service, which will connect up an API and a database service for me.
Then on the backend I can add in the Textract service which can extract text from images.
I can use a simple storage service to store my pictures.
Every time I take a new picture, it is stored, run through Textract and the text content is extracted and stored.
Ok, so now let’s say I want just the information pertinent to investing, and I want that stored and structured nicely.
I add AWS Comprehend which is an AI service that can filter structured text data from text samples, based on pattern identification on some selected key phrases in those samples. Now I have my investing data.
Ok, but I want to make this easy for others too. Instead of having to read the results of the summarize article, I want those results read to them.
Now I add in an Alexa skill, or else I use the Polly service, to transcribe from text to voice, and I send the results to users on their phones.
And so on…
I could keep going with this, pretty much forever. Because all I am doing here is quickly inventing a random scenario, implementing it as a modular patch, then expanding on that invention. And I am just repeating that process over and over.
All of this is possible on AWS, partly because so many services are offered, but also because everything I have described can be done fast. I can, and have, set up working proof of concepts for these things in a matter of hours. And I can keep adding to them and modifying them, just as if I were playing an instrument.
This is why playing the cloud is an apt metaphor. It is like composing results out of thin air. Once you understand this about the cloud, things like ‘on-premise’, ‘hybrid’ and ‘lift and shift’ perhaps won’t look quite so forward thinking as they may have once appeared.
It is just a matter of seeing things differently — or maybe hearing …
Which is good, because one thing that recent events has taught us is that to survive in these times, a business will likely need to quickly adapt and change — and this is what music does too, well, good music, anyway.
OK, But Why AWS In Particular? Surely There Are Other Providers Out There. What About Azure?
I get asked this a lot…
I have three answers, all of them high level, and essentially a matter of temperament.
A: Minor# ‘why do I have to click three times just to achieve this?’ issue
AWS, in common with Apple, make products that just work how I expect them to. Google do this too — but they are not yet in the big cloud players and remain, to my mind, a data company over a cloud company, which is no bad thing.
Another way of putting this is to say that AWS products feel as if they read my mind. They fit with my intuitions, all the time. Microsoft products, sadly, do not. I click something, and another, confusing thing pops up, and then it gives me a message I don’t understand, and so on.
Maybe that’s just me. I am sure Microsoft bring happiness to millions. But I prefer to stick with the stuff that works for the way my mind does.
B: Flat# out side with the innovator over the imitator.
AWS were there first. They invented most of this stuff. I have had to watch, sadly, as Microsoft bring out imitation after imitation of stuff that AWS has already done.
They have every right to do that, of course, and it makes business sense, in some ways. But these companies are bringing out new stuff all the time. So I prefer to side with the innovator. They made the mistakes for themselves and they have the experience to guide them. On top of that, they are more likely to come up with something new and interesting than an imitator.
C: Sharp# er without a safety rope.
In terms of costs and functionality, Azure and AWS are fairly similar today. One of the main reasons people suggest Azure over AWS is that it is well integrated with Windows. And this is true. If you want a ‘cloud for Windows’ you can go with Azure. But this smacks, to me, of Microsoft doing the old trick of leveraging Windows market share to gain share elsewhere. Again, that is fine and makes business sense.
But like Batman escaping the Lazarus pit — one plays sharper without a safety rope.
