As the hybrid cloud continues to build steam, enterprise IT professionals are scrambling to choose a public cloud that is both cost-effective and adequately supports their end users compute requirements,
This, as it turns out, is no easy task. The “Big 3” public cloud providers, namely AWS, Azure, and Google Cloud Platform all offer their own set of unique competitive advantage. That being said, all 3 are not without their drawbacks. Choosing the one that is the best fit for your organization can be dizzying.
Luckily, we can help! Let’s take a closer look at the Big 3.
Amazon Web Services
As the first player in the space, and with by far the largest market share, AWS is usually the first stop when looking for public cloud compute resources. With 16 regional datacenters and four within the United States, there is likely a datacenter near you. Trusted by the likes of Netflix, Unilever, GE, and Kellogg’s, AWS boasts a 99.95% uptime availability. However, in the event of an outage, failover options offer business continuity and peace of mind. In terms of compute resources, AWS is best suited for the cost-conscious looking for CPU-intensive instances.
We like AWS for their “Burstable Performance” Instances. “Bursting” into the public cloud, or in other words, quickly spinning up on-demand instances to meet a temporary need for additional compute is one of the key competitive advantages of a hybrid cloud solution. AWS has built out a pricing model around this idea, so cost analysis of public cloud bursting is relatively straight-forward.
Azure
Azure, formerly known as Microsoft Azure, is Microsoft’s answer to AWS. Azure, although a more recent addition to the public cloud space, is quickly building steam, particularly for GPU-intensive use cases. As well as enterprises that already have a strong relationship with Microsoft.
In terms of availability, Azure is nearly global. Azure is generally available in 36 regions, including two datacenters in South Africa, three in India, and 12 in North America. They do offer some pretty cheap basic instances, but the real selling point is in the Azure N-Series instances.
We love these instances for environments running GPU-intensive applications, such as AutoCAD, Adobe Suite, and AVID Media Composer. Powered by NVIDIA Tesla M60 GPUs, these instances provide nearly bare-metal performance, even when scaling across hundreds of machines.
It’s important to keep in mind that these instances are powerful, and they will cost you. They also require high-performance display protocols, such as HP RGS, and Mechdyne TGX, at an additional cost.
However, for users that require GPU-intensive applications with an “at-desk” experience, the Azure N-Series and a high-performance display protocol is the peanut butter and jelly of graphics-intense cloud computing!
Google Cloud Platform
Not to be outdone, Google launched their public cloud play in 2014 and we’re just starting to see more traction in GPC.
In our research, we found that Google Cloud Platform offers regional availability comparable to AWS and Azure (11 regions and 33 zones within those regions). One thing we found unique about GPC was their custom machine types. Allocate a specific number of vCPU and memory if you can’t find a machine type that is optimal for your workloads. A pretty neat idea!
Don’t be afraid to mix and match!
As the “Big 3” public cloud offerings provide their own set of competitive advantages, we want to stress that you are not limited to just one.
We in the industry talk a lot about hybrid clouds, but less about multi-cloud solutions. A hybrid cloud solution combines on-premise compute, edge compute, and public cloud compute, whereas a multi-cloud solution combines multiple public clouds. Hybrid clouds can be multi-clouds and multi-clouds can be hybrid clouds – what makes the multi-cloud unique is the public cloud mix.