In the UK, large enterprise, household named, logos in all markets are using AWS and not only for virtual machines and storage. Enterprises are consuming features such as identity, data warehouse, search, content delivery as AWS increase the size of the set of undifferentiated heavy lifting due to the hundreds of features they are releasing and improving.
Cloud isn’t a commodity, it’s not about all clouds being the same but rather about them being different. Hybrid cloud between on-premises/cloud is a reality, but cloud-to-cloud could be an unrealized dream because they are fundamentally different (APIs, features etc). Hybrid cloud it’s not a thing or destination but a journey (amen to that!).
Cloud enables experimentation which is a huge business benefit not just for startups. It’s ok for large enterprises to experiment and be lean, in fact they have to do this to avoid being a slow fish that will be eaten by the fast fish.
The enterprise is embracing cloud and there’s no turning back. If you’re not doing cloud, you’re wasting finite resources and making yourself uncompetitive: how do you explain that to the CEO, the shareholders but most important your customers?
Expect exponential take up of AWS cloud services in the coming years to see them eat more of that $3.8Tn enterprise IT spend.
Rather than look at cloud through the enterprise prism; look at the enterprise through the cloud prism to unlock new insights and opportunities.
- Some numbers from the summit
- Werner’s keynote
- Customer stories
- ITV (media platform)
- Omnifone (music platform)
- JustGiving (charity platform)
- GoSquared (customer analytics platform)
- Enterprise Cloud
- Announcements (go see SiliconAngle TV coverage).
Today I had the privilege to be invited to the UK AWS Summit in London. It’s fair to say that after the presentations, meeting the ecosystem and the customers, I’m more impressed with AWS than ever. This isn’t “fanbois fever”: it’s a grown up, considered response to some constant themes that emerged over the day, starting with Eoin Cosgrave, AWS UK & Ireland Regional Manager.
Watch a short twenty minute debrief between myself and Stu Miniman, or read the detail below.
Some numbers from the summit
- 3,000 attendees
- 32 breakouts
- 48 partner exhibits (and some real corkers)
- 22 customers speakers (and big household enterprise names, not just cool startups).
- 4th summit
- 9th birthday for AWS
- Intel were the Platinum sponsor. Think about that.
- S3 usage is best measure by data transfer now, not objects (now it’s in trillions), and growth YoY is 102$ (not including Amazon.com user)
- EC2 usage measure in instances is 93% YoY
- Over 1m active users (past 30 days activity)
- UK household logos in every segment like Transport for London, BP, Shell, Vodafone, St James’ Place, Financial Times, Unilever, Accenture.
- Features released nearly double YoY since 2008 – 24, 48, 61, 82, 159, 280, 516 (2014).
- Redshift datawarehouse is the fastest growing service. Customer feedback has led to 65 feature releases in the past year.
A packed and humid auditorium saw Werner deliver a well-balanced presentation with good mix of numbers, announcements, insights, themes and customer stories.
“Cloud has become the new normal” – when Werner was the CTO at amazon.com he hated dealing with providers because it always felt that they were in charge, expensive. Once the cheque was signed the account manager would disappear. Any changes required costly maintenance agreements. This made AWS want to be the antithesis of those kinds of providers and instead be customer-centric and put the customers in charge, not the IT provider. Obsessing over best service, no contractual obligations, no shackles, no lock in: you are in charge.
This is classic platform play, focusing on demand-side innovation. Read @infoecon and The Second Machine Age for more.
The love affair between startups and AWS is well known because amongst a lot of things, AWS gives these fast moving startups the ability to compete with slow moving incumbents. In the olden days a startup would need $5m capital for IT to get started, which is an expensive experiment, and we all know that startups regularly pivot and will likely have changing IT requirements. What about when they scale up AND DOWN? Look at AirBnB – 400,000 nights a day around the world are brokered by AirBnB and they only have 5 IT staff.
This speed is not just for startups: it applies to companies of all sizes. It is impossible to stay competitive against fast movers and others without using the cloud. Stop wasting time, energy and money in the procurement and refresh cycle. Get rid of capacity planning and sinking cost into peak+15% possible loads.
When Nicolas Carr said “IT doesn’t matter” people often misunderstand him: he wasn’t saying that IT wasn’t needed, he was saying that it wasn’t a differentiator. Why are petrochemical companies building datacenters and racking infrastructure?
Cloud is about more than infrastructure
Household brands are leaning in to cloud not just for virtual machines and storage, but more for data related activities, whether it’s baseball predictions and scenarios with Statcast using missile tracking tech and Kinesis, or Singapore Post launching an ecommerce platform that in two years represented 25% of their revenue.
Customers are consuming more of the services on top of the IaaS layer, such as federated identity and datawarehousing. Lydia of Gartner recently said, “You can’t treat cloud IaaS providers like commodities”. Real world workloads come in all kinds of shapes and sizes and they are consuming all kinds of AWS services that are not available on other cloud platforms. Many other cloud platforms don’t offer the basic IaaS that AWS did in 2010, never mind the plethora of instances (optimized for various workloads) or the additional enterprise features.
Experimentation and mistakes are ok
Cloud gives you the ability to experiment and backout if it doesn’t work without being lumbered with large sunk costs. If you get the workload mapping wrong, you can just change it. 50% of IT projects go over budget and 30% fail entirely because it’s hard to estimate resources (and humans aren’t good at this) and because customers change their minds (which humans are very good at).
It’s the combination of cloud features that make AWS attractive and sticky. The integration of AWS cloud services to form a business process is not possible in other clouds because they don’t have the range of features. If you’re just buying compute and storage then you still have a lot of heavy lifting to do to add features on top, and now AWS are making that previous value add into undifferentiated heavy lifting: they are eating the stack, the tide is rising and you should float on top of it just like your customers should float on top of you.
Cloud is not a binary decision, it’s a spectrum and a journey. The biggest nod to hybrid cloud, Werner acknowledged the on-premises capabilities and how aspects can be merged such as identity, networking, management and deployment. Some companies go “all in on AWS” and others have a mix, and that’s ok. Hybrid cloud is a journey not a destination.
There is a new golden rule in enterprise IT (and a Wikibon position):
Friends don’t let friends build datacenters
Werner finished with a statement that the cloud is gravity so don’t fight it, but I think he should have shouted out
RESISTANCE IS FUTILE!
UK Customer Stories
There were some excellent presentations by customers in Werner’s keynote to expand on the importance of the features in AWS. None of these talked about virtual machines or storage; they all talked about the higher level features.
JustGiving platform play and their use of analytics, Richard Atkinson did a great job
- 23m users, 164 countries, $3Bn raised, 9 currencies
- Launched 2001
- “no great cause goes unfunded”
- Connect causes to people who care
- Used to be about marathon running, now baking, beard growing etc
- Used to be email based, now social, now matching
- Shares, cares and likes
- What you care about = causes you’re connected to + people you’re connected to
- GiveGraph – who has funded which cause and which fundraiser? Social graph underneath it
- $1.6M a day in 2014 fundraising, showed global map, trusted AWS as they re-engineered into microservices
- A lot of real time analytics
- Big spikes (disasters) can’t predict when or where
- Ice buckets, selfies, disasters
- Huddersfield marathon band
Omnifone music platform play, Phil Sant the founder / chief engineer, they do music platform heavy lifting on top of AWS cloud heavy lifting so Spotify and others don’t have to.
- Music as a service
- Streaming download background raido, all audio types
- Old eays 1000sqft dc with colossal architecture
- Customers demanded resiliance, $15m sunk cost in washington to have second DCs
- When millions of users, challenging and expensive, keeping features behaving with consistent performance
- Decided to move to the cloud, AWS only choice
- Needed super high service levels, building the plane in flight
- ALL IN
- Decommissioning datacenters
- Focus on differentiators and not managing infrastructure
- See AWS as a rising tide of features and functionalities, sit like a boat on top, floating and leave amazon to do the UNDIFFERENTIATED heavy lifting
- Their customers let music cloud do the undifferentiated music industry stuff
- ALL IN but they also use SaaS services
- Months in due diligence in s/w suppliers, but now test them and be prepared to switch out for cheaper / more efficient
- Those that don’t move to the cloud will be out-innovated by those that do
- Technology disruption on steroids
- Music industry – fifteen years music companies built their own infra but now renting it from music cloud
Paul Clark of ITV media platform talked about their move from monolithic to microservices, APIs and automation with staggering results.
- Largest free to air commercial broadcaster in UK 2.5bn quid rev, 99.8% of …
- Previous monolithic app stack, poor infra visibility
- New stack metadata, licenses, assets, schedule apis
- Api management, release, monitoring, message
- Automate everything to drive up quality and speed up releases. Automate your testing, build, deployments and infra creation
- Want customers to have a great experience
- Snowflake servers – look the same, but slightly different, wasted time finding problems.
- Months lost waiting for hand built servers, networking
- 21 days to make 1 fw change, we had over 200 to make
- Scripted infra into a new vpc 4 weeks -> 2 weeks, > 28 mins to deploy the whole infra from scratch
- Open sourced tooling, smart caching to protect customers from failure
- ITV Player o new stack in 2014, best quality stats of any ITV player platform?
Geoff Wagstaff of GoSquared was the startup story of how 14yr olds built an online business, not possible without the cloud. They use lots of services like DynamoDB and ElasticSearch.
- London startup, mission is to provide cross platform customer analytics to online businesses to understand their customers better
- 10s of thousands of customers
- Issues for online businesses – customer interaction via mobile apps, CRMs etc
- Create segments and cohorts
- If customers cancel the account, alert/account management
- Thousands of business from Jaguar …. To small
- They were 14 years old when they started1
- Didn’t have time and bandwidth or money to build servers
- Innovations like RDS meant even less to worry about, and benefit from reduced ops
- Lowered barriers, reduced ops, faster deploys
- APIs great for integrating workflows, with CI
- Scale – billions of datapoints per day, don’t worry about capacity planning
- Experimentation – dynamoDB, EMR, cassandra, elasticsearch. They weren’t sure which to use, test on a use case basis
- Removed hurdles, flexible, cost effective – wouldn’t have been able to build business without AWS
- Better company – lowered barriers, faster development
It was excellent to see a whole track dedicated to the enterprise with St James Place financial services being a customer presentation.
Stephen Orban runs the Enterprise Strategy for AWS and has a great background as a CIO of US household brands that moved from non-cloud to cloud and he presented many useful approaches for cloud adoption.
Key: Cloud Adoption Framework.
This is reminiscent, to me, of the frameworks we did at VMware in the early days to win the enterprises: this is a proven approach and I think this is going to be significant in enterprise adopting AWS at the expense of non-cloud and other cloud providers. A brief outline:
- Mapping the Journey to the Cloud
- CAF Perspectives
- Business Perspective: Obtaining Value from the AWS Cloud
- Platform Perspective: Architecting and Designing for the Cloud
- Maturity Perspective: Assessment of Cloud Maturity and Readiness
- People Perspective: Staffing of Cloud IT Teams
- Process Perspective: IT Lifecycle for the Cloud
- Operations Perspective: Efficient IT Operations on the Cloud
- Security Perspective: Achieving Risk, Security and Compliance Goals
There was nothing new since last week’s announcements in San Francisco, so if you want the detail on the announcements then pop along to watch #theCUBE coverage from last week. Below is AWS Chief Data Scientist Matt Wood on theCUBE at AWS Summit SF.