Oser Communications Group

Super Computer Show Daily Nov 18, 2014

Issue link: http://osercommunicationsgroup.uberflip.com/i/412884

Contents of this Issue

Navigation

Page 0 of 11

An interview with Michael Dexter, CTO, Gainframe. SCSD: What product or service does Gainframe provide? MD: Gainframe is a BSD Unix support and custom software provider specializing in FreeNAS, pfSense, FreeBSD and OpenBSD with strong ties to their respective project communities. SCSD: What brings you to SC14? MD: This is our sixth Supercomputing conference, and I absolutely love the science over sales attitude of the event. All of the top vendors are here, but their booths are staffed with highly competent engineers and high-level sales people. Each year I part- ner with a different open source organization to give the open source community the voice it deserves at the event, given the key role that open source plays in HPC. This year, we are hosting the bhyve Hypervisor booth in conjunction with the Open Source Initiative, a.k.a. opensource.org. bhyve is a new hypervisor that appeared in FreeBSD 10, and I am very excited about its potential in conjunction with other FreeBSD tech- nologies like OpenZFS. SCSD: What are FreeNAS and pfSense? MD: FreeNAS and pfSense are open source appliance operating systems based on FreeBSD that provide NAS and firewall functionality, respectively. In my opinion, they are the best GAINFRAME HOSTS BHYVE HYPERVISOR AT SC14 By Rod Wilson, Senior Director, External Research, Ciena Scientific breakthrough is made possible through informed discussion and debate. It is not predicated through knowledge hoarding but rather sharing, which validates the ideas of an individual. This fact leads one to understand the importance of connecting the world's leading researchers and educators and the incredible value it can bring to humanity in general. Recent advancements in big data collection, analytics and its storage have provid- ed the research and education (R&E) community with information that once seemed unattainable, and are leading researchers to groundbreaking scientific discoveries. However, the usefulness of big data is dependent on the end-user's ability to access and share it, as the research community is inherently collaborative. Without an effective means to share all this newly discovered information, the information itself becomes cumbersome and can actually act as an obstacle that researchers must deal with. Luckily for us, R&E networks are already providing the bridge needed between individuals in the research community to ensure collaboration is made possible. For the R&E community, the network is as important a piece of infrastructure as the highways researchers and educators use to get to work in the morning. The Square Kilometre Array (SKA) radio telescope project is a great example of a sci- ence collaboration bound to create incredible amounts of data that requires a reliable con- nection to an R&E network. The SKA project will detect electromagnetic radiation at wave- lengths from meters to centimetres with a collecting area of one million square metres (one square kilometre); 100 times larger than our most powerful existing radio telescope array. By Trond Smestad, President, Numascale LLC Innovative developers can now access the power of shared memory systems for the price point and ease-of-use of a cluster by utilizing Numascale's NumaConnect, a simple add-on card for commodity servers. The hard- ware is now deployed in systems with more than 5000 cores, and the memory address- ing capability is virtually unlimited. The big differentiator with NumaConnect, compared to other high-speed intercon- nect technologies, is its shared memory and cache coherency. These features allow pro- grams to access any memory location and any memory mapped I/O device in a multi- processor system with a high degree of efficiency. They provide scalable systems with a unified programming model that stays the same from the small multi-core machines used in laptops and desktops to the largest imaginable single system-image machines that may contain thousands of processors. The architecture is commonly classified as ccNuma (or Numa) but the interconnect system can alternatively be used as a low latency clustering interconnect. Numascale systems are deployed by simply installing a card with a PCI form fac- tor into a standard server. This approach makes it possible to take advantage of the great price break offered by mass-produced servers with volume applications outside the segment covered by NumaConnect. Servers from IBM, Supermicro and Dell pro- vide excellent building blocks for large memory systems in combination with NumaConnect cards. The design is implemented in a chip, the NumaChip, with an external cache in DRAM, the NumaCache. The NumaChip can address up to 4,095 nodes in a single An interview with Eva Cherry, President and CEO of Silicon Mechanics. SCSD: To what do you attribute Silicon Mechanics' fairly explosive growth over the last several years? EC: Customer focus – without a doubt. We take on smaller projects with the same kind of dedication to a personalized customer experience as we do for large ones. Over the last several years, we have been focusing on build- ing a solid foundation, and we now have the ability to leverage that foundation. Given that, we felt that we owed it to ourselves and our customers to grow our business and increase value for customers and shareholders at the same time. We also worked hard to create and maintain a company culture that attracts top-level talent, since our rate of growth is directly related to the rate at which we are able to attract the right people. Finally and perhaps most importantly, we have practiced organic, profitable growth, which is harder than growing with venture capital or by going into debt, but it is the most sustainable growth a company can have. SCSD: As we near the end of 2014, can you speak to what will be some of the main focuses for Silicon Mechanics in 2015? EC: We will continue to focus on our guiding principle of "Expert Included," striv- ing to understand the individual needs of our customers and personalize their expe- rience with us in terms of reliability, convenience, relevance and responsiveness. As Continued on Page 9 Continued on Page 9 Continued on Page 9 Continued on Page 9 FACILITATING COLLABORATION THROUGH THE NETWORK NUMASCALE PROVIDES PLUG-AND-PLAY SMP – SHARED MEMORY AT A CLUSTER PRICE FOCUS ON ORGANIC GROWTH A GUIDING PRINCIPLE FOR SILICON MECHANICS O s e r C o m m u n i c a t i o n s G ro u p N e w O rl e a n s Tu e s d a y, N o ve m b e r 1 8 , 2 0 1 4 AN INDEPENDENT PUBLICATION NOT AFFILIATED WITH SC

Articles in this issue

Links on this page

view archives of Oser Communications Group - Super Computer Show Daily Nov 18, 2014