Thoughts from the AWS Summit London

Last week, the 10th May, the AWS Summit rolled into town at the Excel conference center in London. I’ve been to every one of these since they started, indeed, as I used to work for AWS. I remember the first of these was me with three other AWS folk, in a room with Werner Vogels, with around 15 customers. How it’s changed….

Thousands descended upon East London to hear the latest and greatest from the cloud giant. Here are my observations.

  1. It all starts with serverless. It didn’t go unnoticed by many that when Werner hit the stage for the keynote, he leapt straight into serverless and his example, iRobot, of a completely serverless architecture. AWS have a strong start in the serverless space, with their flagship FaaS service, Lambda, being the most talked about. They are keen to remind us that they have been delivering serverless services since nearly day 1, in the form of S3, DynamoDB and Aurora, to name a few.

Functionality such as the multi-master for Aurora was a welcome announcement, and it feels like enterprise grade relational databases, as a serverless service, are finally here. As ever, AWS were quick to compare themselves with other vendors who make claims in this space, quickly rubbishing them as being false promises, and architecturally, you can see why they feel right to say this.

  1. Whilst serverless got the keynote going, the substance really arrived when we heard about machine learning. This part of the talk lasted the longest compared to time given to other product areas. We definitely left feeling that AWS are focused on ML enablement for all of their customers, offering a wide range of primitives such as TensorFlow, Caffe, MXNet, CNTK, Pytorch, as well as tools such as Keras and Gluon.

If you don’t want to get your hands so dirty, then you can use the machine learning service to host the compute and algorithyms, allowing you to focus on producing models.

Lastly, if you’re a web developer and just need some ML capability, then they have their suite of services such as Amazon Polly (text to speech), Amazon Lex (speech to text), etc.

Whilst AWS have an impressive array of tools that can enable a range of skill sets to exploit ML capabilities, there is one big area still not addressed: adversarial data. Whilst the services are mature, allowing you to process the data, how can you trust the data? What if someone else injects ‘bad’ data into your learning data, altering the models and producing a negative outcome?

AWS are not alone in this space, and I see little commercially happening to solve this problem. Could blockchain be the answer? Possibly, and there are several efforts being worked on right now, but I’ve not seen anything from AWS in this space — neither a solid blockchain proposition nor the emergence of a solution to adversarial data.

  1. A bit of the same-old same-old. If you had been in Las Vegas for Re:Invent, then there would have been little need to attend this conference. There was very little new, or specific to the UK audience, that you couldn’t have gotten online. Don’t get me wrong — the breakout sessions are always useful if you’re coming at one of the services for the first time, but the recorded sessions on YouTube from Re:Invent cover a broader range of subjects and go into greater degrees of depth.

I attended the Sumerian session, which was a great mix of overview, live demo and deep dive Q&A. Indeed, it inspired me so much I went home and built my first AR scene and found it incredibly easy to do so.

The ecosystem feels a bit stale, with the usual mix of third parties, but as ever, a confusing mix of organisations. Some I found to be totally relevant to the AWS story, But some I felt were at the wrong conference, and their relevance only went as far as providing sponsorship money, which allows the conference to remain free of charge to all who attend. I spoke to several others who felt the need to see something new in regards to ecosystem participants, as well as more companies who are totally relevant.

All in all, it was a good conference, but dare I say it: I was left feeling like I neededto see something new, not from an AWS service catalogue perspective, but more from an ecosystem perspective.

This post first appeared in Glen Robinsons’ blog.


Working for DXC Technology’s Leading Edge Forum, Glen Robinson helps technology leaders develop a Matrix mindset, and the ability to understand the impact of technology for both opportunity and threat. He is an Emerging Technology Director at LEF and a Distinguished Engineer. Read more about Glen and connect with him at @GlenPRobinson.

 

 

RELATED LINKS

How serverless computing serves the enterprise

Getting TensorFlow running on a MacBook Pro

 

 

Comments

  1. Ronald Sonntag says:

    Thank you. Excellent summary.

    Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

w

Connecting to %s

%d bloggers like this: