How Sainsbury’s uses AWS

On Sainsbury’s move and use of AWS, serverless, and DevOps:

“Our relationship with AWS really kicked off at the point we decided to take our groceries online business and rebuild it in the cloud. This was effectively taking a WebSphere e-commerce monolith with an Oracle RAC database, and moving it, and modularising it, and putting it into AWS,” Sainsbury’s CIO Phil Jordan told the audience.

“That movement of RAC to RDS and that big database migration was all done using AWS services, and now we have a fully fledged cloud-native-ish service that runs groceries online across all of our business. Today, we run about 80 per cent of our groceries online with EC2, and 20 per cent is serverless.”

In total, the company migrated more than 7TB of data into the cloud. As a result, or so Jordan claimed, the mart spends 30 per cent less on infrastructure, and regularly sees a 70-80 per cent improvement in performance of interactions on the website and batch processing. So far, there’s been no “major” outages, said the CIO, without defining “major”.

Moving to the cloud has also helped Sainsbury’s into the warm infinity-looped embrace of DevOps. The company has moved from five to six releases per year to multiple releases per day, said the CIO.

Source: Holy high street, Sainsbury’s! Have you forgotten Bezos’ bunch are the competition?

Check out their talk, scrub in to about 24:10.

Related, the Sainsbury’s tech blog is pretty good.

And, from elsewhere and unrelated to Sainsbury’s, some clearer notion that “serverless” forces an event-driven architecture:

So why can’t we just write an event-driven system for our corporate infrastructure? Our world, is event-driven, and generally, we reduce the complexity of our systems by just defining events. “When there’s an access to the FTP service of upload … do this …”, “When there’s an access on a column on a database … do this “. In an IoT world, with billions of disparate devices, it is the only way to go. And if we are to create truly citizen-focused systems, we need to define the events which trigger. How many organisations could crisply define the operation of their infrastructure and all the interactions that happen?

Rather than just defining a server running Exchange, we could have some code which triggers on “When Bob logs-in open up his mail box”, or “When Alice changes the marks for her students, send an update to the exams office”. This is a world where the complexity of servers moves us towards “The Cloud” as a computation resource. In this way we write rules based on events and enact them in the Cloud. There’s no concept of running Exchange or Web servers.

Link: Comic Relief switched from multi-cloud to serverless with AWS and saw a 93% cost reduction

As a team, going serverless has given us a lot more velocity, we can rapidly release, we can test the same infrastructure we’re deploying in production, in a pull request environment, in a staging environment, we can rapidly retest ideas- and every developer can do that because we’re using Lambda to load test, so the power it gives you as a developer and engineering team is pretty amazing.

Source: Comic Relief switched from multi-cloud to serverless with AWS and saw a 93% cost reduction

Link: Serverless survey

“Serverless is growing, and fast. Several key adoption metrics are 2x what they were last year. And not just with smaller companies; the enterprise is adopting serverless technologies for critical workloads just as rapidly.”
Original source: Serverless survey

Link: Serverless Impacts on Business, Process and Culture

‘Sharples said the main interest stems from an enterprise love of microservices, where incremental delivery, agility and faster delivery are being embraced. “But we see adopters struggle with the operational complexity of managing and monitoring distributed systems, and that is where serverless has gotten their attention. You get the microservices benefits, but from a developer perspective it is very easy — it is just about the code. And on the ops side, serverless is a very good model for those building automated ops systems. To respond to a log event, doing that as a serverless invocation is pretty convenient.” Sharples says that in his weekly and daily chats with enterprise customers, serverless is piquing their interest.’

Microservicoed fine in dev, SRE problem now.
Original source: Serverless Impacts on Business, Process and Culture

Link: Serverless Architectures: A Paradigm Shift in …

“One of the biggest security upsides to developing on serverless architectures is that organizations don’t have to deal with the daunting task of having to constantly apply security patches for the underlying operating system. These tasks are now in the domain of the serverless architecture provider.”

The rest – the application code – still needs to be secure. Of course.
Original source: Serverless Architectures: A Paradigm Shift in …

Link: Serverless at Bustle

‘Probably the biggest is: how do you deal with the migration of legacy things? At Bustle we ended up mostly re-architecting our entire platform around serverless, and so that’s one option, but certainly not available to everybody. But even then, the first time we launched a serverless service, we brought down all of our Redis instances — because Lambda spun up all these containers and we hit connection limits that you would never expect to hit in a normal app.

‘So if you’ve got something sitting on a mainframe somewhere that is used to only having 20 connections and then you moved over some upstream service to Lambda and suddenly it has 10,000 connections instead of 20. You’ve got a problem. If you’ve bought into service-oriented architecture as a whole over the last four or five years, then you might have a better time, because you can say “Well, all these things do is talk to each other via an API, so we can replace a single service with serverless functions.”’
Original source: Serverless at Bustle

Link: Finally, a more coherent IBM story?

‘IBM is pushing their cloud hard than ever before. Even though it is mostly IBM Cloud Private, this is the first time I heard a more consistent hybrid cloud story. They have moved away from OpenStack (completely) and CloudFoundry (for the most part) and building a good hybrid cloud story using Kubernetes. Both he product team and services team seem to be quite excited about the opportunity Kubernetes offers them. Kubernetes on Bare Metal as a service is a step in the right direction. Unlike OpenStack and, to some extent, CloudFoundry, Kubernetes gives IBM a chance to have a more unified story from their infrastructure assets to middleware assets. Their “One Cloud Architecture” push is directed in this direction and helps IBM tell a consistent story across their entire portfolio.’
Original source: Finally, a more coherent IBM story?

Link: Finally, a more coherent IBM story?

‘IBM is pushing their cloud hard than ever before. Even though it is mostly IBM Cloud Private, this is the first time I heard a more consistent hybrid cloud story. They have moved away from OpenStack (completely) and CloudFoundry (for the most part) and building a good hybrid cloud story using Kubernetes. Both he product team and services team seem to be quite excited about the opportunity Kubernetes offers them. Kubernetes on Bare Metal as a service is a step in the right direction. Unlike OpenStack and, to some extent, CloudFoundry, Kubernetes gives IBM a chance to have a more unified story from their infrastructure assets to middleware assets. Their “One Cloud Architecture” push is directed in this direction and helps IBM tell a consistent story across their entire portfolio.’
Original source: Finally, a more coherent IBM story?