Cray, an HPE company, held a panel discussion webinar on October 18 to discuss Exascale (10^18, get it?) supercomputing. This is definitely not in my area of expertise, but it is certainly interesting.
Following is information I gleaned from links they sent to me. Basically, it is Why Supercomputing. And not only computers, but also networking to support them.
Today’s science, technology, and big data questions are bigger, more complex, and more urgent than ever. Answering them demands an entirely new approach to computing. Meet the next era of supercomputing. Code-named Shasta, this system is our most significant technology advancement in decades. With it, we’re introducing revolutionary capabilities for revolutionary questions. Shasta is the next era of supercomputing for your next era of science, discovery, and achievement.
WHY SUPERCOMPUTING IS CHANGING
The kinds of questions being asked today have created a sea-change in supercomputing. Increasingly, high-performance computing systems need to be able to handle massive converged modeling, simulation, AI, and analytics workloads.
With these needs driving science and technology, the next generation of supercomputing will be characterized by exascale performance, data-centric workloads and diversification of processor architectures.
Shasta is that entirely new design. We’ve created it from the ground up to address today’s diversifying needs.
Built to be data-centric, it runs diverse workloads all at the same time. Hardware and software innovations tackle system bottlenecks, manageability, and job completion issues that emerge or grow when core counts increase, compute node architectures proliferate, and workflows expand to incorporate AI at scale.
It eliminates the distinction between clusters and supercomputers with a single new system architecture, enabling a choice of computational infrastructure without tradeoffs. And it allows for mixing and matching multiple processor and accelerator architectures with support for our
new Cray-designed and developed interconnect we call Slingshot.
Slingshot is our new high-speed, purpose-built supercomputing interconnect. It’s our eighth generation of scalable HPC network. In earlier Cray designs, we pioneered the use of adaptive routing, pioneered the design of high-radix switch architectures, and invented a new low-diameter system topology, the dragonfly.
Slingshot breaks new ground again. It features Ethernet capability, advanced adaptive routing, first-of-a-kind congestion control, and sophisticated quality-of-service capabilities. Support for both IP-routed and remote memory operations broadens the range of applications beyond traditional modeling and simulation.
Quality-of-service and novel congestion management features limit the impact to critical workloads from other applications, system services, I/O traffic, or co-tenant workloads. Reduction in the network diameter from five hops (in the current Cray XCTM generation) to three reduces cost, latency, and power while improving sustained bandwidth and reliability.
FLEXIBILITY AND TCO
As your workloads rapidly evolve, the ability to choose your architecture becomes critical. With Shasta, you can incorporate any silicon processing choice — or a heterogenous mix — with a single management and application development infrastructure. Flex from single to multi-socket nodes, GPUs, FPGAs, and other processing options that may emerge, such as AI-specialized accelerators.
Designed for a decade or more of work, Shasta also eliminates the need for frequent, expensive upgrades, giving you exceptionally low total
cost of ownership. With its software architecture you can deploy a workflow and management environment in a single system, regardless of packaging.
Shasta packaging comes in two options: a 19” air- or liquid-cooled, standard datacenter rack and a high-density, liquid-cooled rack designed to take 64 compute blades with multiple processors per blade.
Additionally, Shasta supports processors well over 500 watts, eliminating the need to do forklift upgrades of system infrastructure to accommodate higher-power processors.
OPC UA and TSN (Time Sensitive Network). A marriage I was beginning to think was never going to happen. I wrote a preliminary white paper following Hannover Messe 2017. Yes, more than a year ago. (Check it out by clicking the small ad on the sidebar.) This thing has been like a ball in a Rugby match—kicked, going different directions, downed and picked up. People wanting to move before thinking. Getting caught up in legal issues and “politics.” Postponed press conferences.
And, now…”The OPC Foundation launches an initiative to further enable OPC UA adoption throughout industrial automation by extending standardization and harmonization activities for OPC UA including TSN-enabled Ethernet networks.”
The goal of this initiative is to deliver an open, cohesive approach to implement OPC UA including TSN and associated application profiles. This will advance the OPC Foundation providing vendor independent end-to-end interoperability into field level devices for all relevant industry automation use-cases. The OPC Foundation vision of becoming the worldwide industrial interoperability standard is advanced by integrating field devices and the shop floor.
A new set of working groups will identify, manage and standardize the OPC UA relevant topics focused on industrial automation including,
• harmonization and standardization of application profiles e.g. IO, motion control, safety, system redundancy
• standardization of OPC UA information models for field level devices in offline e.g. device description and online e.g. diagnostics
• mapping of OPC UA application profiles related to real-time operations on ethernet networks including TSN
• definition of certification procedures
The working groups will closely align with the TSN Profile for Industrial Automation (TSN-IA-Profile) which will be standardized by the IEC/IEEE 60802 standardization group. This will help ensure that a single, converged TSN network approach is maintained so that OPC UA can share one common multi-vendor TSN network infrastructure together with other applications.
This initiative integrates well with existing joint working groups engaged in ongoing companion specification e.g. description of machines.
Stefan Hoppe, President of the OPC Foundation said “The benefit of membership in the OPC Foundation allows companies to actively engage and influence the direction of the OPC Foundation and includes early access to the specifications and technology. This initiative will grow OPC UA into new markets and I highly encourage all OPC Foundation members to contact the OPC Foundation to participate”.
Thomas Burke, Strategic Marketing Officer of the OPC Foundation, “We are very excited about the initiative to extend OPC UA including TSN down to the field level, and the number of companies that want to actively participate in this initiative bringing the technology into real world products. This set of working groups will pave the way for the broadest, easiest, and fastest market adoption of OPC UA over TSN.”
The OPC Foundation develops and maintains OPC UA as an open and secure communication platform comprised of an information model framework, communication models and underlying protocol bindings. As such, the OPC Foundation works non-exclusively with other organizations on various OPC UA related topics but continues to operate as a platform, technology, use case, and vendor agnostic standardization body.
The ODVA held its Industry Conference and 19th Annual meeting this week in Atlanta. Perhaps more than the Common Industrial Protocol (CIP), the topic of conversation was Michael—the hurricane. We started getting rain from its outer bands in the late afternoon Wednesday. By the time I awoke at 5:30 am to get ready to catch my flight out, it was all over. Not so fortunate were the millions directly impacted. My prayers go out to them.
I have missed the last two or three of these. It was good to get an update. There was no announcement while I was there, but there were some people from Honeywell Process Solutions present who talked about using EtherNet/IP for process automation applications. These switching industry alignments are fascinating to watch. Foundation Fieldbus seems to have lost momentum recently. Will EtherNet/IP, the CIP network, absorb some of the market share?
A well organized series of speakers started Wednesday morning tech sessions with a quick update from all of the SIGs. There are many volunteers putting out an incredible amount of hours developing and updating the various specifications. I can‘t report on them all here—it would be too deep into the weeds anyway. But let‘s just say that ODVA is alive and well.
As even casual readers here know, I am a fan of Time Sensitive Networking (TSN). Yes, I know that it‘s not ready for prime time, yet. Products are beginning to appear in the market, and interest is building across the industry.
I sat in two sessions focused on TSN and CIP. There is technical work going on. The sessions and ensuing questions laid bare the engineering challenges involved in developing CIP over TSN. It‘s non-trivial, but doable. Some may still question TSN, but I‘m even more bullish.
On another front, work has begun on updating the ODVA product data sheet specifications. This work will eventually provide for more and better information to users.
A small group of companies proposed a marketing initiative promoting OPC UA over a new Ethernet standard called Time Sensitive Networking (TSN) in 2017 at Hannover Messe. I was privileged to sit in a meeting to listen to the proposal and subsequently wrote a white paper about it. I believe this is revolutionary technology for the information part of manufacturing technology.
Meanwhile, Rockwell Automation is beginning to regularly surprise me. They first went out of their twice to talk about truly adopting OPC UA and introduced a module for its control platform using it. The company has a long standing reputation for getting involved in standards it doesn’t directly control for the purpose of delaying adoption. But this seemed like a genuine adoption of interoperability recognizing that customers are demanding freely flowing information from a variety of sources.Just to add to my surprise was an announcement I heard about at Hannover that Rockwell Automation has joined that group of OPC UA over TSN companies, now dubbed the “Shapers”. This group is rapidly moving toward critical mass with rumors swirling about companies not (yet) a part of it.
The press release (I haven’t yet had an interview) states Rockwell Automation is joining industry leaders ABB, Belden, Bosch Rexroth, B&R, Cisco, Hilscher, KUKA, National Instruments, Parker Hannifin, Phoenix Contact, Pilz, Schneider Electric, TTTech and WAGO (collectively known as Shapers) to create a communication solution for real-time and sensor-to-cloud applications in industrial operations.
The solution will be based on the OPC UA protocol, which allows easy and secure sharing of information across different vendor technologies and the time-sensitive networking (TSN) suite of standards, which helps improve latency and robustness in converged industrial networks.
“Connecting technologies across an industrial organization while maintaining multivendor interoperability requires a harmonized, interoperable solution that uses consistent information models, communication and application behavior (together known as application profiles),” said Paul Brooks, business development manager, Rockwell Automation.
“That’s what this group of automation leaders are combining their expertise to create. Our solution will give manufacturing and industrial organizations best-of-breed I/O device control, motion and safety application profiles,” said Sebastian Sachse, B&R Industrial Automation.
To ensure the emerging OPC UA TSN solution supports interoperability of different vendor technologies on the same network, the companies are engaging with industry consortia such as Avnu, IEEE, IIC, LNI 4.0 and OPC Foundation. The companies are also planning an announcement in the coming months on how to achieve unified application profiles, which is the last hurdle to device harmonization. They aim to provide one-stop-shop certification of the overall solution up to the device-profile level.
The companies have already published whitepapers on OPC UA TSN technology, such as an IIC whitepaper on converged traffic types. They have also made significant contributions to the recently released PubSub extension of OPC UA, and plan to set up a collaboration between the IIC and LNI testbeds.
This potentially holds great promise for end user companies and systems integrators. We can only hope it progresses.
Ever wonder about the need for the elusive IT/OT convergence? Rockwell Automation announces Factory Talk Network Manager software for its Stratix line of managed (Ethernet) switches. Rockwell OEMs switches from Cisco built to its specifications. Cisco builds good equipment, but it is famous in the networking world for somewhat, shall we say, complex management software.
Control engineers and plant-floor technicians who have growing Ethernet networks to connect all this Internet of Things stuff need something that is closer to their language.
By the way, I still have plenty of catching up to do with things I learned both at Hannover (where I spent many hours with Hewlett Packard Enterprise) and the following week at Dell Technologies World in Las Vegas. I’m finally home and getting organized.
This new management software enables engineers and technicians to monitor the health of their Allen-Bradley Stratix managed switches, troubleshoot switch issues, and quickly configure new managed switches all from one easy-to-use software interface.
“Many plant-floor personnel struggle to piece together information about managed switches and devices from different sources,” said Lorenzo Majewski, product manager, Rockwell Automation. “With the FactoryTalk Network Manager software, they can access this information in one collective spot. In addition, real-time alarms and events from network switches can help them conduct faster, more precise troubleshooting.”
FactoryTalk Network Manager software automatically discovers assets, their associated IP addresses, and creates a topology of these connected devices. The software’s intuitive interface offers grouping of equipment along with dashboard information, so users can organize devices into specific areas or analyze them individually.
The software also uses user-created configuration templates to get new switches up and running faster and more efficiently. These templates can be shared across an organization, or with OEMs and system integrators to further ease network deployments, commissions and maintenance efforts.
The FactoryTalk Network Manager software provides role-based access control with auditing capabilities to help track user-specific activities and changes. The software supports multiple protocols, including Simple Network Management Protocol (SNMP), Common Industrial Protocol (CIP), Modbus, BACnet and PROFINET. Access to the web-based platform is available via a personal computer in a control room or a mobile device on the plant floor.