As we move into 2018, here at Rescale we’re excited to keep our momentum as one of Silicon Valley’s fastest growing enterprise software companies, and more specifically keep advancing the state of the art for HPC in the cloud.
Our product team was busy in 2017. Last year we added the following to our platform: 25 new architectures including Skylake, access to the newest NVIDIA GPU offerings P100 and V100, new bare metal providers and integration partners. Enhancing our HPC workflow support, we introduced reusable clusters and in-browser SSH, added support for large DOEs, and made significant optimization framework and ANSYS RSM improvements. Licensing-wise, we added on-demand, pay-per-use licensing for ANSYS and for LS-DYNA, provided license availability-based queuing for STAR-CCM+ and ANSYS, and introduced license server access automation (via our new startup program). We introduced in-browser desktops for native cloud pre- and post-processing. For enterprise administrators, we introduced deposit pricing support, expanded spot pricing regions, and increased administrative control of workflows, budgets, and data limits. To keep your data secure, we also added advanced connection security and outbound internet blocking for companies. And finally, we provided early access to Intel’s latest architectures via our ScaleX Labs.
And there’s more coming. In 2018, whether you’re an engineer who needs HPC or you manage your organization’s IT and HPC resources, we have something in the pipeline for you. You can look forward to seeing a number of new features and improvements in the areas of cloud robustness, application software, licensing, desktop and visualization, HPC workflows, enterprise organization administration, security and compliance, data transfer and management, and hardware. Keep reading below part 1 of a two-part blog series that will give you a deeper dive into our roadmap in each of these areas.
Rescale makes it ridiculously easy to spin up a high-performance cluster in the cloud so you can start running jobs to completion almost immediately. Spinning up a high-performance cluster directly on the public cloud is a complex task. After navigating all the steps, it takes only one missed configuration option to get a sub-performant cluster. But that’s not all—even after you’ve successfully deployed hundreds of high-powered nodes in the same placement group (to minimize internode latency) and your multi-hour (or multi-day) job is successfully processing, one or more of those nodes may fail. Whether that’s due to hardware issues or spot market preemption, these are issues public cloud providers don’t include in their SLA, but Rescale transparently handles for you.
Specifically, our Spot market offerings are constantly being tuned and optimized to provide you the lowest cost with highest SLA possible. In addition, to ensure the successful completion of your job and minimize your jobs’ time-to-results, in 2018 we will deliver a framework for stalled job and slow job identification and automatic checkpoint restart.
One of Rescale’s greatest features is the hundreds of HPC applications we provide and support. We currently deploy more than 280 ported and tuned applications on our platform and are always adding more (and additional versions of each). Last year we added 80 new applications and 443 new versions, and we’re on track to increase those numbers considerably in 2018.
In addition, we’re enhancing your ability to deploy your own applications on Rescale, so you can provision your own software in multiple cloud providers with a minimum of clicks.
Finally, we are very proud of our highly-skilled engineering staff, who embody thousands of hours of deep experience in each of the applications we support and are directly available via live chat through our ScaleX platform. In 2018, we’re growing our team to deliver the best support possible.
Licensing is one of the biggest challenges of HPC in the cloud. HPC applications are very complicated, representing a huge development effort by ISVs, and they command correspondingly significant licensing fees. Rescale transparently accesses on-premises license servers for cloud HPC jobs, giving enterprises the flexibility to use the same license server for both their on-premises and cloud jobs. In addition, Rescale provides on-demand licenses for many of the HPC software applications we provide, for example ANSYS Elastic Units. In 2018, expect us to add many additional HPC software applications to our on-demand offerings, as well as integrated viewing, administration, and reporting on your company’s license utilization in on-premises, multi-region, and even multi-cloud environments.
That’s all for now! Check back in a week for a rundown on the rest of our product roadmap. In the meantime, if you’re interested in becoming an early adopter for any of our updates mentioned above, please don’t hesitate to contact us directly at email@example.com and we will make sure you are one of the first to be able to give it a try!
This article was written by Dave Anselmi.