Aussie Broadband prioritises VMware-based workloads for cloud migration

Aussie Broadband has completed the initial tranche of its internal cloud build and is progressively modernising workloads to run on it.

Head of cloud and SRE Ben O’Shea told the iTnews Podcast that the cloud platform, built with hardware and software from Cisco, NetApp, Pure Storage and SUSE, is live and hosting some infrastructure support systems so far.

“[These are] systems that drive or orchestrate our [infrastructure] automations and stuff like that,” O’Shea said.

“Those are running on there at the moment, and we're working with our different business stakeholders [now] to identify their applications and how we prioritise and sequence those [for modernisation and migration to the new internal cloud environment].”

When the internal cloud platform was first revealed, one of its stated aims was to replace Aussie Broadband’s existing “corporate virtualisation platform” and existing “Docker and Kubernetes platforms” for hosting workloads.

O’Shea said that when it comes to workload migration, the company is currently prioritising those that run in VMware-based virtual machines.

“Exiting our internal VMware platforms is a priority for us,” he said.

“We're wanting to avoid renewals [or] upgrades of a platform that we're not going to invest in going forward.”

VMware’s absence from the new internal cloud technology stack was a deliberate design decision.

“We looked at our requirements when we were doing our internal cloud platform, and I guess there were just a few things playing out in that VMware-Broadcom space that didn't really sit well with us in terms of values alignment … [around] Broadcom and their approach to their customers and the technology industry in general.

“Being a telco, being in a commodity business, managing our underlying costs is imperative to staying competitive and being able to deliver the required shareholder returns and margins that we need to deliver.

“We also are leaning heavily towards containerisation rather than traditional virtual machines virtualisation. VMware is definitely not a market leader in that space.”

O’Shea said the company would take a risk-based approach to application modernisation and migration once the VMware exit is complete.

“What is the oldest workload? What's giving us headaches from a compliance perspective? Those will be prioritised, and we'll be looking at opportunities to modernise some of those going forward where it makes sense.”

Backup change

The company has also changed the way it backs up applications and workloads hosted both on its new internal cloud and older VMware and Docker/Kubernetes environments, switching to Veeam.

"We had a different backup platform," O'Shea said.

"We'd gone down a particular road and we weren't seeing the performance at scale that we needed to see with that backup [solution]."

O'Shea said that Aussie Broadband was "supported really well by Veeam to run a small scale proof-of-concept", initially on its "legacy platforms", which showed it could meet Aussie Broadband's requirements.

"We now run not only our next generation cloud platform on that Veeam solution, but also the legacy Kubernetes platforms and the legacy environments are protected by it as well."

Formalising SRE

Since he first appeared at Cisco Live to reveal the cloud platform build, O’Shea’s role has evolved from general manager of transformation and cloud platform, to general manager of cloud and SRE.

This coincided with the launch of a new business strategy at Aussie Broadband called ‘Look to 28’, which encompasses a “technology transformation … to establish common core systems and processes” across the group.

O’Shea retains his responsibility for cloud - both internal and customer-facing services - and gains formal responsibility for “modern site reliability engineering practices”.

“We brought in an SRE leader to establish our SRE capability, and we've got some roles in-market as well to grow that function,” O’Shea said.

“We're currently looking at internal transfers as well, and that person and the team that we're starting to slowly build out is meeting with all of our key technology and business stakeholders, understanding what their reliability needs are. 

“We've also launched an internal community of practice where people who are interested in availability and how we can improve availability and reliability can come together and share ideas and challenges for increasing the awareness of that internally, and that's helping us to promote the value that SRE can bring.”