How to Summit AWS — an Architect’s perspective

By Martyn S

TL;DR — AWS Summit Online is a learning Summit; I Learnt a lot about how to kickstart our cloud adoption; Stretch before you do 5 hours of video talks!

Hey, I’m Martyn. I’m a Solutions Architect working for Gemba Advantage in London. I’ve worked with Public Sector customers for over 13 years and have a passion for web apps, elegant user interfaces and keeping it simple!

Personally I never got around to attending a tech Summit before, and people who work in tech will find that strange. The truth is that it was never something that I myself, or my employers encouraged. I started a new role in April with a fresh attitude after all these years. It’s important to invest in yourself and I’m happy to be part of something different now at Gemba Advantage.

This Wednesday was my chance, although things are a bit different as instead of visiting the ExCeL London, I’d be attending from my office chair. I’m sure we’ve all found it challenging to sit in front of a screen for 5 hours, but I really wanted to absorb as much content as possible and that was my goal.

Of all the changes I’ve had to make now I’m working from home, the most valuable? Stretching. A few shoulder rolls, a bit of twisting side to side, and some reaching above your head does it for me. My warm-up for the day was literally a warm-up.

Why I promoted it to my teams

My new role as of April, has me in a position of responsibility. Responsible not only for my personal development as before, but responsible for technical direction and growth of 5 teams of engineers.

We have recently started focussing on the cloud, and are kickstarting our cloud adoption with attendance encouraged for this event.

AWS Summit Online is billed as a Learning Conference, so I figured that this was the perfect opportunity for the teams to take a leap forward in understanding of what cloud actually is, and how to build with cloud in mind.

Keynote 09.15–10.00

The Keynote resonated with me as a Solutions Architect, with interesting introductions to Builders Library, Well-architected Framework, and interesting presentations from startups that are based entirely in the cloud.

Part of how I approach Solutions Architecture is to research best practice and pull from different patterns and approaches to similar problems. Looks like Amazon are trying to reduce the need for Solutions Architects like myself, by training their “builders” in Solutions, providing tools like the Well-architected-tool to assess implemented solutions, and AWS Solutions Library as a resource for best-practice Architectures.

Dr. Werner Vogels also spoke about how load balancing and content delivery services are still used as foundational building blocks and support auto-scaling at potentially massive scale.

Amazon seem to be focussing more on training and certification, and now provide 360+ free online courses, along with Q&A style videos, and blogs targeted at all levels of engineer.

Choose your track

First job is to “choose your track”.

I actually did this the day before the Summit. In my excitement I actually printed the “Agenda at a glance” sheet and went at it with a red pen, circling the talks that sounded interesting to me. This proved mostly successful, but after we “wrapped-up” the day with the Gemba team, some different and interesting approaches to choosing talks came out.

My personal approach was to circle anything that I felt could change my thinking on topics that I already knew about. I didn’t consider the “Level” at the bottom of each option denoting the experience/knowledge required to understand each talk. A risky strategy given my fairly limited experience in some areas of AWS and I ended up with circles across 7 of the 11 tracks, and 10 circles in total. Ultimately I attended talks across 3 different tracks and managed to attend 4 sessions in total.

A couple of colleagues in senior positions chose a track and stuck to it. One, with deep experience in implementing DevOps went for the “I’m in DevOps”. The other, wanted to learn more about how AWS is having an impact in our sector, so went for “I’m in public sector”. Both had completely different Summit experiences as the DevOps track was full of Level 300 talks with serious technical depth, whilst the public sector track had nothing above Level 200 with one particular talk providing very little value to this individual.

An engineer I work with ended up inadvertently sticking to what he knows in choosing talks on the “I build applications” and “I’m in DevOps” tracks. His driver was to be able to apply what he learned to his near term work packages, and he feels that he was largely successful in achieving this goal.

My colleague who works in business engagement actually looked at the Levels available and selected the “beginner” (Level 100) talks initially in the “I’m new to the cloud” track, later switching to “I’m in public sector” as he wasn’t interested in the pricing talk.

Breakout Sessions 10.00–13:15

As you can see above, there were 55 sessions across the 5 hour period. My general thoughts on the talks that I attended are that they were really fast! Ricardo Suieras seemed to realise that he had to get it done in 30 mins, and caught him smiling as he was reading what I assume was his autocue. I ended up in a couple of his talks, which was surprising given the number of talks and presenters.

I assume that all the talks were pre-recorded as the captions were a line ahead of the audio. I wonder how this works in the real world? Do presenters take questions? Are the talks longer?

The nice thing about the Summit taking place online is that the talks are available on-demand after the event and all the slideware was available to download as PDF.

Raw data to business insights: What you need to build a modern data lake (30 mins)

It was all going so well… We’d heard from Dr. Vogels about AWS’s amazing auto-scaling capability… And then… we couldn’t get into any talks! Ended up getting in after about 5 minutes of refreshing, but certainly not the best advert for AWS.

Javier Ramirez did a good job with the Data Lake talk. Again surprised how much he managed to cram in! Really learnt how much you can do in AWS with a few clicks.

He demoed AWS Lake Formation that uses AWS Glue, and showed how with some raw data loaded into an S3 object store, you can set a Crawler to process it and then query using Amazon Athena, and visualise using Amazon QuickSight. All the services were pay-as-you-go with the exception of the S3 storage (which has it’s own innovative ways to make things cost effective).

It struck me that it’s sometimes not all about code and infrastructure as code, which I’m still pushing engineers to adopt. Feels good to have the safety net of good administrative interfaces especially when considering ops.

Accelerate innovation with hybrid capabilities and infrastructures (30 mins)

This was a really interesting talk about how companies who want to adopt the cloud deal with the challenges of migrating existing workloads, or running them whilst building new capability in the cloud.

Ricardo Suieras talked about how migrating completely takes time so adopting a hybrid cloud architecture can get you there sooner using common services and APIs across workloads.

AWS really do make it easy to adopt their ecosystem. They have VMware migration options to get running on AWS and there seem to be a lot of data migration options such as Data Migration Services (DMS) and datasource connectors for Athena.

Security is a major consideration when connecting on-premise workload to the cloud, and AWS have thought about this too with the use of Virtual Private Gateways (VPNs).

This all had me wondering if it’s this easy to get on AWS, how easy is it to get off it? What if I want to use other could providers in the future in a multi-cloud pattern? I’m directing engineers to be aware of abstraction, and portability not proprietary for our strategic assets. But the benefits of AWS, and its ecosystem for accelerating innovation are there to see.

Application Integration patterns for microservices (30 mins)

Application Integration patterns for (not only) microservices was the best session for me. Anshul Sharma did a great job of explaining the asynchronous messaging architecture patterns and made it feel like it was about the technology and the engineering rather than about selling the participants more AWS services.

He started with “Divide and conquer” and “Loose coupling, not lousy coupling”, highlighting the benefits in terms of the ability to scale, evolve, and operate services independently. And of course the pitfalls of synchronous pattens; tight coupling, downstream dependencies, and amplification of problems.

Next up was lots of asynchronous integration patterns including “one-way”, “request-response”, “point-to-point”, “publish-subscribe”, and my favourite “topic-queue-chaining”

Anshul had promised us some fun earlier in the talk, and this is where he delivered as he put the patterns to work for a fictional unicorn hailing app startup, Wild Rydes! He showed what bad integration would look like and how loose coupling is better than lousy coupling.

Event-driven architecture/Let’s encrypt everything, really everything/I’m done with talks and need some lunch (45 mins)

At this stage I thought I’d try a talk on Event-driven architecture. I just couldn’t get into it. It felt like the presenter was going over things that I already knew about like APIs. I thought it would be a good idea to switch talks as I had circled 3 for this 45 minute period.

So I switched over to “Encrypt everything” and at this point felt like the universe was telling me it was time to have a break. No talk was going to get me motivated at this point, especially one on encryption!

Time to have some lunch and get away from the screen. Turns out 3 hours of screen time without snacks and coffee is my limit.

Migrate your enterprise applications to the cloud (30 mins)

Re-energised after lunch I actually got a lot out of Ricardo’s second talk of the day. I was really interested to hear Amazon’s recommended approach to migrating enterprises to the cloud; their process is Assess, Mobilise, Migrate & Modernise.

I’d heard I lot about the 3rd stage, Migration in the previous talk on hybrid cloud, and I’m actually going through a similar process of assessment with my customers. I was able to validate that my approach of starting with training and skills development, was a good one.

Included in the Assess stage was information about readiness assessments, and Amazon even provide the AWS Cloud Adoption Readiness Tool to help generate a readiness report https://cart.splashthat.com/. I’ll be looking to see how this works for my customer.

Next Ricardo talked about the Mobilise stage where he described the need to plan at setup and consider the following:

There are lots of different migration options for your Windows capabilities including lift and shift for Windows Servers nearing the end of support, and licensing options using pay-as-you-go (PAYG).

Amazon give you the opportunity to reduce costs by migrating to open source, moving from .NET to .NET Core, using MS SQL Server for Linux, or using their DMS (Data Migration Service) to move to a new open source DB platform. They also offer an innovative approach to compute sizing, reducing the number of cores in a cluster deployment to make licensing that little bit cheaper; caveated by the fact that the reduced core compute is the same price as the full offering.

I was happy to see a slide with the 7 R’s (formally the 6 R’s) that I have been using on the products in my engineering organisation to determine the right strategy for migrating to the cloud.

The final stage is Migrate and Modernise. There was some overlap with the previous talk on Hybrid Cloud talk although the AWS Migration Hub was revealed in this talk. A tools to track the migration status of your applications, across multiple tools in a single place.

Closing Remarks 13.15–13.45

The closing remarks were the standard sales job that I’m informed the big tech companies end their events with. Felt like this was more geared towards company CEO/CTOs, than my teams of engineers.

Wrap up

Cloud adoption using AWS is an important thread that runs through our company, and the work that we do for our customers. At the end of the day we got together as a company to “wrap up” the day and talk about our experiences.

The size and scale of this online Summit obviously caused some issues. Attendees had trouble joining talks, and the quality of certain streams was poor. I do remember at one stage, I had to download the PDF of the slides, and open them up alongside the video stream so I could make out the detail. On the positive side, a colleague of mine did comment that, the online version of this event seemed more “laid back” as when it’s hosted at the ExCeL, it can be “competitive” to get into the more popular breakout sessions.

Technical issues aside, all the feedback that I have received about the event has been positive. People managed to attend multiple talks, and learned a lot of new things about AWS. The greatest value seems to be when people attended talks that aligned with their existing knowledge. Providing them with a different perspective, a cloud focussed perspective. Exactly what I wanted for my teams when attending the Summit.

I’ll certainly be attending more of these types of event in the future, and I’ll be applying the new techniques I’ve learned. I’ll be using AWS services to accelerate innovation and experimentation, following the principles of hybrid cloud, I’ll be figuring out how ready my organisation is for the future, with the Cloud Adoption Readiness Tool and report, and I will certainly make sure to continue to stretch before multiple hours of screen time!