From the blog

Unlocking HPC Cloud Transformation for Enterprises

Today I get to share something exciting – something that I believe will be a true enabler for the world of applied science and engineering.  

I joined Rescale in part because I strongly related to its mission to accelerate science and engineering innovation by transforming high performance computing (HPC). From battling pandemics, to addressing climate change, solving the world’s hardest science and engineering challenges require HPC. It’s also everywhere – HPC enables nearly every physical product we use today, from the safer cars we drive, to the advanced smartphones we use.

Given HPC’s pivotal role enabling businesses to commercialize new product innovations, it’s striking how slightly  this ~$40B market has changed in recent years. Only 20% of HPC workloads are in the cloud, compared to ~80% for general enterprise workloads. Today, all major cloud service providers (CSPs) offer specialized HPC infrastructures – great progress, but challenges remain.

The first challenge was turning raw cloud HPC infrastructure into turnkey simulation platforms that engineers find useful – something Rescale is really good at.  During my travels late last year to meet and listen to our customers, one comment stood out.  “Rescale, you guys make cloud tolerable,” said a highly experienced HPC practitioner.

He explained that in his previous job he managed fixed on-premises HPC infrastructures solving for high utilization.  But now he works at an engineering-driven company, where his primary responsibility is maximizing engineering throughput.  This meant using the cloud (which has the necessary scale and latest HPC architectures), providing a broad simulation software portfolio, staying on budget, and doing it all securely in a highly regulated industry.  According to him, without Rescale’s automation platform, cloud HPC expertise, and support, this simply would not have been possible. A win for their engineering team.  

But there’s another challenge that we need to tackle, and that’s helping engineering teams and enterprise IT work together.  Tension between IT and lines of business (LOBs) is an age-old problem not unique to HPC. IT is charged with an enterprise-wide security, reliability, and efficiency strategy, while LOBs pursue business outcomes. Until recently, most enterprise digital transformation efforts ignored engineering workloads.  But this is changing: HPC spend growth rate is twice that of overall IT, and business leaders are starting to pay attention. This is an unprecedented challenge for HPC service owners as the economics of HPC have been historically opaque.

As a former consultant serving business technology leaders, I’ve seen this movie before. In helping IT organizations improve their interactions with the business, the solution typically included four things: data-driven transparency, clear rules of engagement, a service catalog for the business to make cost/performance tradeoffs, and continuous improvement. It’s time to apply this discipline to HPC, particularly since cloud computing is an enterprise-wide CTO & CIO decision.

And so Rescale Insight was created, to deliver data-driven business management for HPC.

And we’re just getting started.  Today we also announced the industry’s first AI engine to match simulation workloads with the best architecture from our multi-cloud infrastructure library. We’re putting our eight years of cloud HPC experience into platform intelligence. Now in tech preview, Rescale’s coretype AI gives simulation engineers the best of AWS, Azure, and GCP without changing their workflow in any way.  All this happens within IT-defined policies on which cloud providers, price points, or pre-defined architectures are enabled. 

Rescale Insight provides enterprise IT transparency and control, while empowering engineering teams to drive new product innovations. Simply put, it is a bridge between enterprise IT and LOB engineering teams so that organizations can focus on why they do HPC in the first place:

To accelerate science and engineering breakthroughs.

To bring new product innovations to market.

To change the world.

So, today, I’m excited to finally share Rescale Insight, the new standard for enterprise HPC business management.

Related articles

複数ジョブ実行時の時間と経費を節約する新機能:Persistent Clusters (永続クラスター)

RescaleのRahul Vergheseが、2017年1月19日に記載したBlog記事の翻訳です。 元記事はIntroducing Persistent Clusters: a New Feature to Save Time & Money with Multiple Jobsをご覧ください。 Rescaleは、最新のデプロイメントで新機能のPersistent Clusters (以下、「永続クラスター」:マニュアルで起動/削除可能なクラスター)をリリースしました。この機能を有効にすることで、複数のインスタンスを起動してクラスタを構築し、シャットダウン(訳注:Rescaleシステムではシャットダウン後インスタンスは削除されます)することなく、Rescaleワークフロー(Web UI)を使用して複数のジョブを順番に同じクラスターへ投入できます。以前は、各ジョブ毎にクラスターが稼働し、ジョブの完了後は自動的にシャットダウンされるため、複数の小さなジョブを実行すると遅延が発生する可能性がありました。この新しい機能により、繰り返し処理の高速化が可能になります。これは、テストや同じハードウェア構成を必要とする複数のジョブに特に便利です。 時間とお金を節約 一般に、各クラスターがスピンアップしてシャットダウンするまでには数分かかります。永続的クラスターを有効にしておくと、クラスターに追加する各ジョブの時間とコストを節約できます。 なぜ? 標準のクラスターは、ジョブが完了すると自動的にシャットダウンし、後続のジョブも同じように起動してシャットダウンするため、別々のクラスターとしてそれぞれ課金されます。(訳注:通常、たとえ10分の計算であっても1時間分の課金となるため、10分で完了する連続する2つのジョブを実行した場合、2時間分が課金されます)一方で、永続クラスターを使用すると、クラスターはすぐに次のジョブの実行に使用できるようになるため、ジョブ間で別のクラスターをシャットダウンして起動させる時間を無駄にしません。それによって、同様のジョブを多数立ち上げるユーザーにとって、時間とコストを大幅に節約することになります。(訳注:上記の例だとちょうど10分の計算を待ち時間なく2つ連続的に実施できることになり、1時間分の課金で収まることになります。)

read more »

Rescale Attends Largest Oil and Gas Technical Conference

Last month Rescale attended the Society of Petroleum Engineers’ (SPE) Annual Technical Conference and Exhibition (ATCE), the largest technical conference in the field of oil and gas exploration and production (E&P). This year it was held in New Orleans, LA […]

read more »

Leave a Reply

Your email address will not be published. Required fields are marked *