AWS continues to court enterprise customers with infrastructure features and tools, providing EC2 features that...
enhance instance management.
One feature, EC2 Fleet, addresses cluster management challenges, with the ability to define target capacity, instance types and pricing models. AWS spins up and optimizes the infrastructure. EC2 Fleet resembles AWS' prior Spot Fleets offering, but extends the capability to On-Demand and Reserved Instances -- or any mix of the three models. Customers can also spin up instances in different availability zones.
Previously, enterprise customers had to run scripts or workflow engines through a service like CloudFormation, Elastic Beanstalk or OpsWorks, but EC2 Fleet could make orchestration a more seamless process.
"[EC2 Fleet] lets you specify the higher-level configuration of what you want your fleet to look like," said Deepak Mohan, analyst at IDC. "Fleets will choose the best set of instances available, depending on what you specified as your fleet deployment, and choose the instance types and constitution of the mix for you, instead of you having to do it yourself."
The feature could appeal to a variety of enterprises with sizable batch-processing workloads, such as banks and pharmaceutical and insurance companies, said Carl Brooks, analyst at 451 Research. "It simplifies and streamlines routine deployments and removes a handful of essentially menial tasks in running a limited-term AWS environment," Brooks said in an email.
Avoid extra costs
AWS also recently unveiled Optimize CPUs for Amazon EC2, a feature that lets users specify the number of virtual CPUs (vCPUs) they may want when they spin up instances. By pairing cores with instances, a bring-your-own-license customer might be able to avoid extra fees. The Optimize CPUs feature could also help customers who run Oracle on AWS, as Oracle had changed its licensing in a way that made it costlier to run Oracle software in the public cloud.
"[This could] possibly be a long-term defense against the per-vCPU billing announced by Oracle," said IDC's Mohan, who added that he has yet to see Oracle actually enforce this higher pricing model on its customers.
Additionally, the EC2 feature enables a user to disable hyper-threading, which can impair high-performing computing workloads or legacy app migrations to the cloud.
Lastly, AWS disclosed a 15% price cut on its H1 storage-optimized instances. The cut, Mohan said, lowers the adoption barrier for potential big data or disaster recovery workloads, though the tradeoff comes with heavier data gravity and potentially higher storage costs.
Enterprises drive VM focus
The recent double-down on infrastructure and EC2 features shows, as much as AWS wants to tout advanced services in AI or serverless, enterprises still move to the cloud in deliberate fashion. While Lambda provides a sort of connective tissue between services, it's not necessarily a game-changer for cloud newcomers.
Enterprises may plot grand designs, but their initial moves tend to be conservative, if they do anything at all. "Most enterprises have a goal, at some point, to put stuff in Lambda, in containers and to move to a scale-up architecture. But the fact remains that in 95% of the cases, enterprises are moving workloads that are running on prem into the cloud into an environment that is as similar as possible to their on-prem environment," Mohan said.
AWS' recent focus on EC2 features reflects the needs of those all-too-important enterprise customers, who likely helped boost the cloud provider's growth rate in the first quarter of the year.
"This ties back in with who is moving into the cloud now and how much of the growth in cloud is being driven by enterprises," Mohan said.