by Sachin Murdeshwar

Mumbai: LINODE, accelerates innovation by making cloud computing simple, accessible, and affordable to all. Founded in 2003, Linode helped pioneer the cloud computing industry and is today the largest independent open cloud provider in the world. Headquartered in Philadelphia’s Old City, the company empowers more than a million developers, startups, and businesses across its global network of 11 data centers.

Alex Pena, Lead Research & Development Engineer, Linode on GPU Trends and more In an Interview with APN

1. What recent trends have you witnessed and how do you see an uptick in the demand for cloud-based GPUs in India? Can you share some use cases?

We see a trend in the industry where people want to do faster and parallelized workloads, which is a perfect use case for GPUs (graphics processing units). Once confined to personal computers and gaming rigs, cloud-based GPUs are now growing in popularity for enterprise-grade computing and accelerating highly customized applications and business workloads in the cloud. There is so much that you can do with GPUs between AI, deep learning, and video processing, among others. 

There has also been a very large uptick across geographies as GPUs are pushing technical and creative boundaries in a wide range of industries. For instance, vehicle manufacturers are using GPUs to bring autonomous driving to the masses and -commerce websites are using them to accurately predict buying patterns of shoppers. On the other hand, universities and research groups are using GPUs to simulate physics, solve complex problems, and understand genes to find vaccines for new viruses and diseases, including COVID-19, and media and entertainment companies are relying on GPUs to create visually-rich feature films through simulation, real-time rendering, and virtual production.

2. How do you compete against AWS, Google and Microsoft? Can you highlight some of the GPU instances that you have?

Linode is unique in the sense that it’s one of the first Alternative Cloud providers to offer GPU compute to the masses. In fact, Cloud Spectator recently conducted a study to evaluate the performance of GPU-based virtual machines (VMs) from Linode, Amazon Web Services (AWS), Microsoft Azure, and Google Compute Engine (GCE) against a wide range of workloads. Linode beat the next fastest VM by about 9%, doing this at a significantly lower price point than the other cloud service providers. The RTX 6000 VMs from Linode outperform the competition when it comes to rendering workloads. In addition to our performance, we are far more competitive on a pricing standpoint with the same stack of services and features than the big three. Another big differentiator for us is our customer support. When you call us, you’re not getting a pre-recorded voice message that says ‘press one for this, or press two for this’. There is an actual person on the other end who will troubleshoot your problems and walk you through the solutions.

3. The Cloud Spectator’s report ran a suite of benchmarks, including video and image rendering, can you give us a sense of the price performance advantages that emerged from the report? 

By using benchmarks on multiple GPU VMs, Cloud Spectator found that Linode offers excellent value when it comes to GPU-based VMs. Linode far exceeded the other cloud service providers in terms of the GPU performance per dollar spent. The primary drivers for our strong price-performance are a competitive pricing strategy, which includes simple and straightforward pricing model across all data center locations as well as low price points in general, especially for hourly on-demand pricing. Our smallest plan comes with 32GB of system memory, 8 CPU cores, a RTX 6000, and 16TB of transfer included in the $1000 per month price. With other providers, pricing can be much more complicated, sometimes based on each data centre location. 

4. What challenges are Linode’s GPUs helping users overcome?

In my opinion, GPU instances is a growing market in India, and the first challenge that we users face is in terms of awareness, especially among MSMEs and start-ups that want to run an AI or deep learning or some other neural network. And sometimes they are cash strapped and cannot throw away money at hardware or have the expertise necessary to put it into a rack in the data center. So that is one area where we are helping users with our round the clock customer support and of course, lower price points.

The other challenges that we are helping users overcome are limitless compute on demand which enables them to engage compute power for as little as an hour—or persistently for months on end—which is often cheaper than purchasing and maintaining hardware on premise. We are also offering simplified IT management, which helps users remove on premise maintenance and management barriers, allowing them to focus on their core competencies. And lastly, easy updates, which lets them reap benefits of the latest hardware and software to support a wide range of GPU use cases in the cloud.