Liftr Look Ahead: Liftr Cloud Regions Map Updates - Accelerator Brands
The Liftr Cloud Regions Map is a visual tool provided by Liftr Insights allowing investors to see up-to-date cloud components supplier availability (processor and accelerator brands) and other key infrastructure-as-a-service (IaaS) and platform-as-a-service (PaaS) offerings in every region of each tracked cloud service provider (CSP) – AWS, Azure, GCP and Alibaba Cloud.
Many of the pull-down options in the Liftr Cloud Regions Map are derived from data from the Liftr Cloud Components Tracker. Think of the Liftr Cloud Components Tracker like measuring share of shelf in a supermarket or other retail storefront – it informs customers about the complexity of a tracked cloud’s IaaS offering. Parts of that complexity are the share of services that use different processor and accelerator chips, including the versions of each chip, as well as how much memory is used and more.
Let’s look at a couple of examples showing how the “accelerator brands” pull down option is useful to cloud investors.
We’ll start out on the US West Coast in Silicon Valley. Alibaba Cloud, Azure and AWS have regions located directly in Silicon Valley. According to the Liftr Cloud Regions Map:
- Alibaba Cloud’s us-west-1 region has AMD FirePro S7150 and NVIDIA Tesla P100 GPU accelerators;
- Azure’s us-west region has NVIDIA Tesla K80 and M60 GPU accelerators;
- And AWS’ us-west-1 region has Xilinx Virtex UltraScale+ VU9P FPGA accelerators, plus NVIDIA GRID K520 and Tesla M60 and T4 GPU accelerators.
With the exception of AWS’s NVIDIA Tesla T4 GPU accelerators, these CSPs are offering fairly old accelerator product in Silicon Valley.
Now, let’s move over to Richmond, Virginia:
- Alibaba Cloud’s us-east-1 region offers NVIDIA Tesla T4, P100, and V100 GPU accelerators;
- Azure’s usgov-virginia offers NVIDIA Tesla K80, M60 and V100 GPU accelerators;
- and AWS’ us-east-1 region offers Xilinx Virtex UltraScale+ FPGA accelerators, NVIDIA GRID K520, Tesla K80, M60, T4 and V100 GPU accelerators, AWS’s disaggregated Elastic Grahics and Elastic Inference GPU accelerators and AWS’s own in-house designed Inferentia deep learning inferencing accelerator.
NVIDIA Tesla T4 and V100 GPU accelerators are very current, along with AWS’s recently deployed Inferentia inferencing accelerator. Market demand for public cloud-based deep learning accelerators in Richmond appears to be much greater than in Silicon Valley. This might be a statement about Silicon Valley companies using on prem deep learning gear, or it might be an indicator of a large number of defense and government contractors located in the Richmond area.
The Liftr Cloud Regions Map can also show where hardware is not deployed. Mousing over to Salt Lake City, we can see that Azure’s us-west-central region doesn’t offer any accelerators. The same goes for AWS’ us-gov-west-1 region and Azure’s canada-east region in Quebec City.
While our Liftr Cloud Regions Map is great for interactively checking what is available now in individual regions, Liftr Cloud Components Tracker customers receive a detailed summary of month-to-month deployment changes as they happen. For example, tracking the latest accelerator chip deployments every month enables investors to cross-check otherwise hard to validate success stories from the accelerator chip vendors earnings reports and press releases.
Certainly, knowing how many new deployments of latest generation accelerators by which CSPs in what regions is a great way to check up on how well accelerator vendors are competing against each other.
Make sure to also check out our Liftr Cloud Regions Map videos on processor brands and CMGR.
Liftr Cloud Components Tracker: http://bit.ly/2QceXlT
Liftr Cloud Regions Map: http://bit.ly/2LGB5PV
Follow Us On: