Servers: The Right Compute
cancel
Showing results for 
Search instead for 
Did you mean: 

4 Ways the New Compute Experience Delivers More Agility

Krista_S

How can you have the best possible agility on premises? Let HPE automate the manual work, freeing up resources while helping get the most out of apps and data with predictable performance and intelligent automation. 

Blog_agile_compute.jpgA few years ago, when set out to design our next generation of HPE compute, we asked ourselves a number of key questions which are even more pertinent today.

How can we have the best possible agility on premises?

How can we improve security to handle evolving threats?

How can we help our customers control their economies of scale and react to unpredictable demand?

Let’s focus on tackling that first question: How can we have the best possible agility on premises? The word Agility says it all—things need to be done very fast. The pace of change shows no signs of slowing. IT has moved from being a service to the core of what your company is able to achieve in terms of innovation, new business models and efficiencies. It is all about managing data that is being created everywhere. Getting insights from that data is the new gold.

The first challenge when you run on-premises infrastructure is the overhead—a lot of manual work that often requires special skill sets. Our goal at HPE is to automate as much if not all this manual work, freeing up your resources and allowing IT generalists to perform tasks that previously required special skill sets that have become hard to find or retain in recent years. We can help you achieve these goals in four ways:

  1. HPE Integrated Lights Out (iLO): HPE iLO is the industry’s leading server management solution, two decades running. It’s a solution that differentiates itself by enabling three crucial tenets of server management: configuration, monitoring and remote management. The new version iLO 5 comes with several improvements to simplify operations such as the automation of maintenance upgrades. This dramatically minimizes the work required to perform these operations at scale.
  2. HPE Scalable Persistent Memory: We see the need for speed with data-centric applications. We are building on our in-memory compute innovation with HPE Scalable Persistent Memory—a huge enabler for fast data-centric applications. You can now do much larger in-memory compute with persistence (terabytes instead of low 100s of GBs with NVDIMMs). With this new technology, you can accelerate applications for fast caching and storage, reduce transaction costs for latency-sensitive workloads, and deploy bigger, more affordable datasets to gain new insights from large memory pools. The performance increases are impressive. One example: We took an in-memory Microsoft Hekaton DB and restarted with SSDs in 20 mins. We took that same Microsoft Hekaton DB and restarted with Scalable Persistent Memory in 45 seconds (yes, you read that correctly!).
  3. HPE OneView 3.1: We know it is difficult to migrate existing applications or stand up new applications. Automation streamlines routine tasks and processes performing what your people do manually today into a template. It embodies all the knowledge of optimizing the deployment we have gained by working with our ISV partners such as Mesosphere, one of the new templates on offer in HPE OneView 3.1. OneView is the infrastructure software layer that is now template driven so that applications can be migrated and deployed automatically.
  4. HPE Intelligent Systems Tuning: Reducing latency particularly with mission-critical applications has never been more important. Through our unique partnership with Intel, we are introducing HPE Intelligent Systems Tuning, which is something quite incredible in terms of managing the clock speed of the CPU based on the workloads that you are running. A few technical terms here. The first is Jitter Smoothing—the ability to smooth the fluctuations in processor frequency as you increase performance through Intel’s Turbo Boost mode. This is immensely important to latency sensitive applications such as high frequency trading. Then comes Core Boosting, where, in working with Intel, we have the unique ability to dynamically increase the turbo frequency. It does involve special SKU’s with thermal innovations but does not void the CPU warranty. And finally, Workload Matching—a menu of profiles so you can match key parameters to a specific workload. We have 15 profiles matching the more common customer workloads such as low latency, mission critical, transactional database and high performance compute. These profiles will automatically match the internal resources to the needs typically required by that workload, imparting the knowledge we have gained over the years from performance benchmark tuning.

Today you are making important decisions on where to locate applications, whether it be in your own data center or somebody else’s. At HPE, we realize the one size does not fit all and we are committed to helping you get the most out of applications and data with predictable performance and intelligent automation.  

Why wait? Here’s why it’s time to demand a new compute experience.

About the Author

Krista_S

Krista Satterthwaite has worked in the server industry for 20 years in product management and marketing roles. She is currently the Vice President of Portfolio Marketing for Hybrid IT.