I spent a chunk of time at Amazon’s AWS Summit in London earlier this week. It was interesting to be back at the ExCeL exhibition and convention centre as the last time I was there was another big vendor event – Microsoft Future Decoded – and on the topic of organisation, let’s just say that whilst Amazon had the registration sorted (Microsoft had some issues with that last year), Amazon’s schedule meant we couldn’t get to the technical tracks because of the queues for the escalator (there were no stairs that I could find!). It seems that ExCeL’s ICC has some logistical challenges when dealing with a few thousand people moving from one level to another!
I was interested to see which partners were exhibiting in the Expo though:
There seem to be plenty of opportunities around cloud services but one set of partners who noticeably absent were the big systems integrators. As we moved to the keynote, I couldn’t tweet the highlights as connectivity became an issue (standard conference issues around WiFi and mobile phone access – plenty of signal but too many sharing it) but I settled into Dr Werner Vogels’ (@Werner) presentation and found it interesting on a number of levels:
- Whilst there were announcements about Amazon releases, many of the points made were equally applicable to other clouds (e.g. I could apply the same issues and learning to Microsoft Azure):
- Start-ups have no legacy/dependencies, a low cost structure and can move quickly to disrupt log-standing industries.
- Not that long ago millions of dollars were required to start an Internet business. It was “indirect funding of the American server industry” – but not any more: cloud infrastructure and platform services have a low barrier to entry and no up-front costs or vendor tie-in.
- The cloud’s not just for start-ups: enterprises can benefit too (although I’d argue that variable costs are a challenge for some finance departments and new features arriving daily are not always a good thing for end users).
- Automation is key [in IaaS/PaaS]: whether it’s for testing, building, deployment or infrastructure creation.
- Real-world workloads come in all shapes and sizes – there’s no need to standardise on the lowest common denominator when you can change the services you use to fit. With cloud you can make mistakes (fail fast and move on) that would be expensive using physical servers or even a virtual infrastructure platform.
- AWS is a platform and an ecosystem – some organisations are creating their own platforms on top of AWS (e.g. Omnifone’s B2B media platform).
- Invention is continuous, with new services, and a movement towards micro-services based on smaller blocks (Dr Vogels used a Tetris analogy), containerisation, event-driven computing, etc.
- Security is a shared responsibility – Amazon provide the tools (and a workbook for compliance with local laws) and customers need to use them correctly.
- There may be compliance benefits from the cloud, for example: if you store customer data then you become a data controller – and if you process it on AWS, Amazon becomes a data processor; by signing the AWS data processing agreement organisations comply with EU data processing requirements known as article 29, providing assurances that are not available with on-premises services.
- The stats involved (and Amazon’s growth) are enormous:
- 102% year on year increase in data in/out of S3.
- 93% year on year increase in usage of EC2.
- Over 1 million active customers (every month – excluding amazon.com) – from start-ups to enterprise, in many vertical markets.
- Five times the compute capacity of the other 14 providers in the Gartner Magic Quadrant for cloud infrastructure as a service combined.
- AWS’ pace of innovation is such that 516 major new features and services were launched in 2014 – almost double the previous year (and that’s a pattern with: 24 in 2008; 48 in 2009; 61 in 2010; 82 in 2011; 159 in 2012; 280 in 2013; 516 in 2014) – it’s actually really hard to keep up (and I’d say the same for Azure too!).
- Whilst Amazon is leading this sector, they did not come across as arrogant – indeed Dr Vogels highlighted that the Amazon motto of being “the earth’s most customer-centric company” applies to AWS too. Customers are in charge, not providers and, if Amazon’s not providing what they need, customers will walk away – perhaps not literally (there’s no lock-in but moving workloads is non-trivial) but they will use another cloud for their next project/programme.
One message that I found difficult to swallow was that “hybrid IT is part of the journey, not the destination“. Maybe that’s true in the long term but, in my world of Microsoft cloud services, I see real use cases for hybrid clouds (I’ll be writing more on that soon).
Maybe it’s because I’m coming from a SaaS angle though, rather than IaaS and PaaS – I did attend a hybrid cloud deep dive session at the AWS Summit but missed the first 10 minutes as I was delayed by the need for micro-escalator-services or the “escalator of things” (credit to Vitor Domingos @VD for that thought).
I can see why, when running your own applications, there is little need for long term usage of on-premises infrastructure or so-called “private cloud” platforms once you’ve taken advantages of the efficiencies that cloud IaaS/PaaS provides. So, creating a hybrid cloud may be a bridge to the future state, rather than a two-way street; but maybe there are limitations about the data that you want to store in the cloud? Perhaps the cloud BC and DR is still too risky (although maybe you could consider multiple clouds)? Or connectivity challenges might exist that rule out exclusive cloud usage? Or perhaps I’m just a Luddite!
One thing’s for sure – applications shouldn’t be moved to the cloud “as is” – cloud migration is an opportunity to rethink how things work – and Amazon has an AWS Cloud Adoption Framework that might help with that.
Unfortunately, I couldn’t stay for the whole day – there were many more technical tracks I wanted to attend (and huge queues to get into the sessions) but I had an appointment elsewhere. The AWS Summit was certainly a worthwhile investment of my time though (even as a Microsoft consultant) – I’ll be watching out for it again next year.
In conclusion, I’ll make just the one observation: systems integrators need to find new opportunities to deliver value in an increasingly agile and commoditised world – for example as cloud integrators – and they need to move quickly. Incidentally, this is not news – but for a while now we’ve been able to say “cloud is not for everyone”. Increasingly though, those barriers to cloud adoption are being broken down.