For more than a decade, Intel has been the industry standard when it comes to data centre processors.
The company’s Xeon range of chips has dominated the market thanks to a record of innovation, strong vendor support and a lack of competition.
Despite the high margins on offer, Intel hasn’t had a serious rival in more than a decade. That is, until AMD decided it wanted a piece of the action.
The EPYC range
The launch of AMD’s EPYC range of server chips back in June 2017 was a significant moment for both the industry and for AMD. It marked the company’s first major chipset targeted at the data centre since the introduction of the Opteron line back in 2003.
Based on AMD’s Zen CPU architecture, the first EPYC chips boasted impressive specs and were pitched as a highly scalable platform for the data centre. This is especially true of the single-socket variety which promises to take up less server space, strong performance and superior energy efficiency.
One of AMD’s advantages is that it has expertise in both CPU and GPU technology – unlike Intel which focuses on CPU – and is therefore capable of leveraging its capabilities and knowledge between the two markets. This integration helps drive innovation across its portfolio.
AMD’s share of the market is still small, but it is growing. Data centre revenue for the most recent quarter beat analyst estimates and the company is targeting a double-digit market share within the next 18-months.
Although Intel remains the dominant player, it will be concerned that there is finally a credible challenger.
A significant cause of optimism within AMD is the arrival of its 7nm EPYC processors. AMD has the next generation ‘Rome’ 7nm EPYC processor-based Zen 2 CPU in the works for mid-2019, along with the third-generation ‘Milan’ chipset afterwards.
Strong industry support
Crucially, AMD has gained significant vendor support. Cisco, Hewlett Packard Enterprise (HPE), Dell EMC and Lenovo, all of whom have released products based on EPYC architecture.
Meanwhile, cloud providers have got on board. In December 2017, Microsoft Azure became the first global cloud customer, while Yahoo! Japan and Baidu have also adopted EPYC technology.
And in March 2019, Amazon Web Services (AWS) detailed two new EC2 services powered by AMD EPYC processors.
What’s clear is AMD’s entry has ignited a market in which competition has been at a premium, offering a differentiated product line that appeals to manufacturers and cloud providers.
But is AMD-powered architecture right for your data centre?
It’s true that the enterprise market tends to adopt new technologies a little later than the cloud industry. And when it comes to a private data centre, organisations can be forgiven for playing it safe.
After all, these are the servers that power critical applications and host important corporate and customer data. Any disruption would be a threat to business continuity.
The importance of these systems is why so many have stayed loyal to Intel for CPU technology. And Intel’s dominance has made the data centre market a difficult sector for any vendor to crack – especially when AMD has been a non-entity for so long.
So, the fact that AMD is even in the conversation is testament to its data centre credentials. Not only is its technology impressive, its momentum in the desktop market and its recent advances in servers show it has staying power and should be considered.
Mix and match workloads
On paper, certain elements of EPYC are superior to Intel’s Xeon chips – especially when it comes to efficiency. For organisations looking to build scalable architecture or with limited rack space, AMD-powered equipment might be the superior option.
But although Intel’s 10nm chips are larger in size, the general consensus is that the company will remain the dominant player for some time to come. Indeed, it has some technological tricks up its sleeve to maintain this position.
For example, Intel is pairing its Xeon processors with its Optane memory technology – something that AMD cannot match. Earlier this year Intel detailed its own ‘Cascade Lake’ scalable data centre platform, along with a range of data centre memory products, that provided a timely reminder that it is more than capable of competing with AMD’s 7nm offering.
Intel’s technological roadmap is evidence of the benefit that competition can bring to a market that may have otherwise stagnated. Competition not only drives innovation, but lowers prices, while organisations and vendors have the ability to mix and match technologies based on the unique requirements of their workloads.
The era of hybrid cloud and multi-cloud has made it easier to optimise applications and workloads, so it is significant that this flexibility has been extended to on-premise deployments. AMD is no flash in the pan and Intel isn’t going anywhere. Organisations now have two viable paths to follow.
Contact one of our experts today to leverage the latest AMD innovations for your business here.