metamorworks -

Amazon brings public cloud services to enterprise datacentres with AWS Outposts

AWS builds out hybrid cloud play by offering enterprises access to its public cloud services in private datacentres, hosted on the same hardware its own facilities use

Amazon Web Services (AWS) is to allow enterprises to run its public cloud services in their private datacentres using the same hardware the firm favours in its own facilities.

From the second half of 2019, enterprises will be able to buy managed, configurable on-premise storage racks from the cloud giant that features AWS compute and storage services, the firm’s CEO Andy Jassy announced at its re:Invent user conference in Las Vegas.

These racks – known as AWS Outposts – can be used by enterprises to run workloads that are not ready to move off-premise yet, while making it easier for enterprises to integrate them with ones that are already running in the AWS public cloud – to create a hybrid cloud environment.

There will be two AWS Outposts available at launch, including a version that will run AWS cloud services natively, while the second looks set to build on the firm’s hybrid cloud-focused technology tie-up with VMware.

As previously reported by Computer Weekly, the two technology firms joined forces in October 2016 to bring their VMware Cloud on AWS offering to market, which is geared towards helping enterprises migrate vSphere-based private cloud workloads to the Amazon public cloud with greater ease.

As a continuation to this, enterprises with large VMware deployments within their datacentres will have the option of opting for a VMware Cloud on AWS Outposts variant.

This version will feature the entire VMware software-defined datacentre (SDDC) compute, storage and networking stack, running on-premise, and can be managed using the VMware Cloud on AWS management console.

Read more about AWS

The AWS-only version, meanwhile, will allow enterprises to apply the same APIs and control panels they use when managing the Amazon cloud deployments in an on-premise context.

When announcing AWS Outposts during the keynote, Jassy said the technology’s aim is to provide enterprises with a consistent user experience regardless of where they choose to run their applications and workloads.

“We have customers with a number of workloads that are going to live on-premise for a very long time. And there are lots of different reasons for that. It’s often because they need really low latency to something that sits on-premise,” said Jassy.

“And what they have asked us for is a way to provide AWS services – like compute and storage – on-premise, but in a way that seamlessly and consistently interacts with the rest of my applications and services in AWS.”

This is an area some of AWS’s competitors, including Microsoft and Google, have attempted to address through the roll-out of their respective Azure Stack and its Cloud Services Platform offerings.

Without naming names, however, Jassy went on to suggest its competitors had struggled to give enterprises the type of user experience they were after.

“Enterprises need it to be the same and consistent,” he said. “They want the same APIs, same control panel, hardware and functionality, and if you look at the options out there today to solve this problem, they’re not the same. It’s why those options aren’t getting much traction.”

Hardware specifications to be decided

During a post-keynote Q&A session with the press at re:Invent, Jassy said the exact hardware specifications for the stack are still to be decided, but it is not the firm’s intention to replicate the entire AWS cloud experience within them.

“It remains to be seen how many total services we will have on Outposts or not. Our initial goal is not to recreate all of AWS in Outposts. They are different delivery models, and folks who have tried to do that today have frustrated customers,” he said.

“There are some really basic components we are hearing consistently that customers would really like us to provide on-premise... [including] compute, storage, database, machine learning and analytics.”

Read more on Infrastructure-as-a-Service (IaaS)

Data Center
Data Management