Alexa

50 Smartest Companies Of The Year 2018

Expand Beyond the Limits with ScaleMP

thesiliconreview.com/story_image_upload/us/thesiliconreview-shai-fultheim-ceo-scalemp-18.

We are the Leaders in Virtualization for High-End Computing, Providing Higher Performance and Lower Total Cost of Ownership.

ScaleMP is the leader in virtualization for in-memory high-end computing, providing higher performance and lower total cost of ownership as compared with traditional shared-memory systems. The company’s innovative Versatile SMP (vSMP) architecture aggregates multiple x86 systems into a single virtual x86 system, delivering an industry-standard, high-end shared-memory computer. Using software to replace custom hardware and components, ScaleMP offers a new, revolutionary computing paradigm.

With installations in more than 30 countries around the world, solutions based on vSMP Foundation provide superior performance for database, analytics and simulation applications in commercial, government, and public sectors. These solutions conserve the advantages of the traditional shared-memory systems operating model (ease of installation and management as well as large memory) while leveraging the cost structure of x86-based servers and providing access to a broad applications portfolio.

ScaleMP has assembled a unique team that spans visionary technologists to seasoned industry veterans and the company’s solutions are recognized worldwide as the affordable alternative to proprietary SMP systems, with over 300 customers worldwide, from Global 2000 firms to top research and academic institutes.

Deep Dive into the Products of ScaleMP

The vSMP Foundation™ product suite aggregates multiple, industry-standard, off-the-shelf x86 servers into one single virtual high-end system.  Hundreds of customers use the vSMP Foundation suite of products as an alternative to traditional, proprietary, and expensive symmetrical multiprocessing (SMP) systems, benefiting from the cost structure of x86 clusters. vSMP Foundation also provides cluster users and administrators with the simplified clustering infrastructure of a single operating system.  vSMP Foundation scales up to 2 PB of memory and up to 32,768 cpus.

Flash Expansion:

vSMP Foundation Flash Expansion aggregates DRAM and NVM memory capacity to a single memory space for use in industry-standard, x86 servers, using advanced virtual machine technology. Organizations that run workloads with medium to high memory requirements can meet these workload requirements using an economically efficient combination of memory technologies. Built on the proven Software Defined Memory capabilities of vSMP Foundation over network fabrics (SDM-F), vSMP Foundation Flash Expansion leverages novel Storage Class Memory (SCM) devices to provide Software Defined Memory over Storage (SDM-S). vSMP Foundation Flash Expansion is used by organizations for the following use cases:

DRAM Replacement: Workloads which can fit into a COTS server system using high-end memory configuration can now be executed with much lower DRAM capacity installed. For example, instead of installing 768GB of DRAM into a dual-socket COTS server, you could install 128GB DRAM and approximately 800GB of high-performing NVM, saving on both CAPEX and OPEX.

DRAM Expansion: Workloads which require more memory than a COTS server system no longer require high-end systems. For example, a dual socket COTS server could now easily reach a total of 6TB system memory. Further, previously-unimaginable opportunities for leveraging in-memory computing are now a reality. Organizations can finally put a 30TB database into system memory and reap the benefits of the faster transaction (OLTP) and analytics (OLAP) processing.

vSMP Foundation

vSMP Foundation™ aggregates multiple, industry-standard, x86 servers into one single virtual high-end system. This flagship product provides customers with an alternative to traditional, expensive multiprocessing (SMP or NUMA) systems and also offers simplified clustering infrastructure with a single operating system.

With vSMP Foundation multiple servers can be consolidated into a single, easy to manage the system. Multiple servers now appear to the administrator, developer, or user as a single entity, running only one copy of the operating system (OS). Only one set of updates, patches, and applications need to be installed, as compared with many when running a cluster of systems.  Each and every CPU in the aggregated system has access to all of the memory.

vSMP Foundation / Memory Expansion gives users the ability to access all of the aggregated memory of multiple systems from a single or multiple threads. By combining a number of servers together in order to expand the amount of directly addressable memory, developers and users can easily access many Terabytes of memory without having to learn complicated programming environments. vSMP Foundation / Memory Expansion gives users a cost-effective system that scales the memory to the desired amounts.

vSMP Foundation Advanced Platform

vSMP Foundation™ Advanced Platform extends the capabilities of vSMP Foundation, providing enhanced capabilities over the traditional SMP systems available with vSMP Foundation.

Like vSMP Foundation, vSMP Foundation Advanced Platform offers on-demand SMP, which provides greater utilization by reusing the same systems for both distributed and shared-memory workloads; partitioning, which enables different groups to work on a dedicated resource rather than share the same system and interfere with one another; and parallel InfiniBand performance, which uses multiple HCAs in parallel to improve performance for your most demanding applications. But it also adds support for very large SMPs, which delivers greater system scalability.

As workloads change, a datacenter with vSMP Foundation Advanced Platform installed can react to these changing workloads. Large, virtual SMP’s can be created as needed for users who need access to hundreds of cores and/or Terabytes of memory.  A virtual SMP can be created on the fly by using a number of popular cluster management applications. When the challenging workload decreases or changes to smaller resource-intensive applications, the servers can be returned to the cluster to continue to do useful work.

Selective Scaling

The vSMP Foundation product suite aggregates multiple, industry-standard x86 servers into a single virtual high-end shared-memory system. Depending on workload requirements, either all of the resources (CPUs, memory, and I/O) can be aggregated, or just the memory of the nodes. We call this “selective scaling”, as the organization can adjust the system exactly to the application needs. This provides customers with an alternative to proprietary, expensive high-end systems. vSMP Foundation provides two aggregation modes:

Memory Expansion: Designed for applications which: (1) require a large memory footprint; and, (2) are not CPU demanding. It enables organizations to use larger system memory for analytics and big-data applications. vSMP Foundation provides Software Defined Memory (SDM) capabilities that enable all-in-memory computing and provide applications with a performance boost while reducing license cost. Memory expansion can be achieved by aggregating DRAM over Fabric (SDM-F) or by aggregating DRAM and high-performance flash storage (SDM-S).

System Expansion: Designed for compute-intensive applications. All of the processors, memory and I/O devices of a collection of x86 nodes are aggregated into one shared-memory system. It provides customers with a superior alternative to traditional, expensive multiprocessing (SMP or NUMA) systems.

Additionally, it offers a simplified clustering infrastructure with a single operating system, to be used by administrators looking to benefit from server consolidation, or to reduce the complexity of cluster management.

vSMP Foundation Free

vSMP Foundation Free allows users to create a powerful virtual large-memory system from industry standard servers. By aggregating a number of servers into one virtual machine, users can get access to very large amounts of memory from any application. Any application can get access to up to 1.5 Terabyte (TB) of memory. For example, users can:

  • Keep an entire database in memory
  • Work on more complete genomics information
  • Simulate larger MCAE models by using more memory for the model and temporary I/O.
  • Run larger research projects with more complete physics
  • Make quicker decisions by investigating more data.  i.e.  Big Data

Solutions Portfolio

thesiliconreview-50-smartest-companies-of-the-year-issue-cover-18

In-memory Computing: ScaleMP delivers the one platform today’s organizations require to address the efficiency and performance challenges associated with managing very large amounts of data. Designed for environments that process complex or unstructured data, or data that requires the use of multiple algorithms, vSMP Foundation is a software-based SMP that holds data and enables different processes to operate on the same data while in memory. By retaining more of the frequently used data in memory, the overall analytics capability and speed to results is increased.

Shared-memory Computing: Proprietary high core-count, large-memory systems are expensive and often underutilized. A software-based SMP, vSMP Foundation™ delivers record-breaking performance at a fraction of the cost of proprietary systems — and is vendor-agnostic. By utilizing industry-standard servers, the fastest processors are available for a shared-memory system, as large as is required.

On-demand Computing: As business demands change, server workload requirements change, and the ability to create the right type of system for maximum productivity is paramount. With vSMP Foundation’s on-demand computing capabilities, large-memory or shared-memory VMs can be created as users’ applications require large amounts of memory or CPUs in a single instance of the OS. These servers can then be returned to the cluster as the workload changes.

Consolidated Computing: vSMP Foundation runs a single OS across the entire system. Using the Versatile SMP™ (vSMP) architecture, this ground-breaking solution aggregates up to 128 servers into a single, virtual x86 machine, simplifying day-to-day administration and eliminating the need for cluster-management skillsets.

Accelerated-Computing: vSMP Foundation supports the latest GPU products from NVIDIA. All attached GPUs will be recognized by vSMP Foundation and are available to all applications running on the system. For example, if 8 systems exist, each with 2 GPUs, then all 16 NVIDIA GPUs are available to the application.

Clientele Reviews

“Thanks to vSMP Foundation memory aggregation technology, we experienced a significant increase in performance for many heavy tasks. We are now able to use a variety of applications on a single system and are able to enhance our research.”

  • Professor Leonid Brodsky, D., Senior Lecturer, Department of Evolutionary and Environmental Biology and Tauber Bioinformatics Research Center, University of Haifa

“vSMP Foundation gives us the ability to run very large scale de novo sequencing applications without having to deal with the complexities of managing a cluster. We get full performance from the servers, and we can scale our applications like never before.”

  • Kazuharu Arakawa, Project Assistant Professor (Bioinformatics, Systems Biology)

“We are very pleased with the scale-up performance of vSMP Foundation software. SAC is able to run a variety of workloads on a single system, enabled by vSMP Foundation software.”

  • Kandarp Pandya, Sci. Engr SF, AIPD/SIPA/SAC

“vSMP Foundation has allowed us to advance our biomedical research in complex disease modeling and diagnostic prediction to more accurately represent the underlying biological disease processes leading to the rapid and accurate identification of patient conditions. These advanced clinical decision support tools are instrumental in saving patient lives.”

  • Jeffrey Ferraro Ph.D., Chief Technology Officer and Senior Informaticist at the Homer Warner Center

The Mastermind behind the Masterstroke

Shai Fultheim, Founder, President, and CEO: As founder and CEO of ScaleMP, Shai designed and architected the core technology behind ScaleMP, and is responsible for the company’s strategy and direction. Shai has more than 20 years of experience in technology and business roles in IT and venture-backed firms. Before founding ScaleMP, Shai was CTO of BRM Capital, a first-tier Israeli venture-capital fund. Shai defined the fund’s technology roadmap, which formed the foundation of BRM’s investment strategy.

Prior to BRM, Shai was co-founder, CTO and VP R&D at several technology startups. He also served in the Israeli Defense Force’s prestigious central intelligence unit, where he led a large IT organization of hundreds of engineers and programmers. Shai managed a broad range of complex IT projects in the areas of security, systems and network infrastructure, and knowledge management, gaining significant experience in the domain of scalable computing and large-scale IT infrastructure.

Shai has been an active member of several open source initiatives such as Apache, Jakarta Tomcat, Amanda and the Linux kernel. Shai holds a B.Tech and BASc from the Jerusalem College of Technology.

“Our founder’s in-depth market knowledge and technology vision, together with the executive team’s experience in bringing to market industry-leading products, as well as the commitment and expertise of its employees, positions ScaleMP as a leader in redefining high-end computing architectures.”