Oracle OpenWorld 2019 is the platform for countless software announcements, but since 2010 the company has been in the hardware business thanks to the Sun Microsystems purchase, and the company remains committed to delivering integrated hardware and software systems.Proving the point, the company took the wraps off the Oracle Exadata X8M designed for acceleration of Oracle’s database applications, featuring new data analytics and business intelligence features along with Oracle's newfound religion on automation.The new Exadata X8M server platform uses second-generation Xeon Scalable processors and Intel's Optane DC persistent memory to accelerate performance. That's a big win for Intel, which is seeing quite a bit of momentum for AMD's Epyc processor. And it's another win for Optane, which pretty much every server vendor supports.To read this article in full, please click here
When Microsoft launched Azure as a cloud-based version of its Windows Server operating system, it didn't make it exclusively Windows. It also included Linux support, and in just a few years, the number of Linux instances now outnumbers Windows instances.It's nice to see Microsoft finally shed that not-invented-here attitude that was so toxic for so long, but the company's latest move is really surprising.Microsoft has partnered with a company called Skytap to offer IBM Power9 instances on its Azure cloud service to run Power-based systems inside of the Azure cloud, which will be offered as Azure virtual machines (VM) along with the Xeon and Epyc server instances that it already offers.To read this article in full, please click here
Dell EMC has updated its PowerMax line of enterprise storage systems to offer Intel’s Optane persistent storage and NVMe-over-Fabric, both of which will give the PowerMax a big boost in performance.Last year, Dell launched the PowerMax line with high-performance storage, specifically targeting industries that need very low latency and high resiliency, such as banking, healthcare, and cloud service providers.The company claims the new PowerMax is the first-to-market with dual port Intel Optane SSDs and the use of storage-class memory (SCM) as persistent storage. The Optane is a new type of non-volatile storage that sits between SSDs and memory. It has the persistence of a SSD but almost the speed of a DRAM. Optane storage also has a ridiculous price tag. For example, a 512 GB stick costs nearly $8,000.To read this article in full, please click here
IBM has rolled out a new generation of mainframes – the z15 – that not only bolsters the speed and power of the Big Iron but promises to integrate hybrid cloud, data privacy and security controls for modern workloads.On the hardware side, the z15 mainframe systems ramp up performance and efficiency. For example IBM claims 14 percent more performance per core, 25 percent more system capacity, 25percent more memory, and 20 percent more I/O connectivity than the previous iteration, the z14 system. [ Check out What is hybrid cloud computing and learn what you need to know about multi-cloud. | Get regularly scheduled insights by signing up for Network World newsletters. ]
IBM also says the system can save customers 50 percent of costs over operating x86-based servers and use 40 percent less power than a comparable x86 server farm. And the z15 has the capacity to handle scalable environments such as supporting 2.4 million Docker containers on a single system.To read this article in full, please click here
Looking to support ever-increasing workloads, Cisco has pumped up the speed and intelligence of its storage area networking family.The company addressed the need for increased speed by saying it would add support for 64Gbps ready SAN fabric across its 9700 line of MDS storage directors.The MDS family includes the 18-slot 9718, the 10-slot 9710 and the six-slot 9706.
More about backup and recovery:
Backup vs. archive: Why it’s important to know the difference
How to pick an off-site data-backup method
Tape vs. disk storage: Why isn’t tape dead yet?
The correct levels of backup save time, bandwidth, space
The main idea here is that customers can upgrade to the new fabric module and upgrade their software to add speed and capacity for high-speed fabrics without having to rip-and-replace any of the directors, according to Adarsh Viswanathan, Cisco product manager, data center switching. A 64Gbps line card for all three chassis will be available in the future, Cisco said.To read this article in full, please click here
AMD's $5.4 billion purchase of ATI Technologies in 2006 seemed like an odd match. Not only were the companies in separate markets, but they were on separate coasts, with ATI in the Toronto, Canada, region and AMD in Sunnyvale, California.They made it work, and arguably it saved AMD from extinction because it was the graphics business that kept the company afloat while the Athlon/Opteron business was going nowhere. There were many quarters where graphics brought in more revenue than CPUs and likely saved the company from bankruptcy.But those days are over, and AMD is once again a highly competitive CPU company, and quarterly sales are getting very close to the $2 billion mark. While the CPU business is on fire, the GPU business continues to do well.To read this article in full, please click here
Tests by the evaluation and testing site ServeTheHome found a server with two AMD Epyc processors can outperform a four-socket Intel system that costs considerably more.If you don’t read ServeTheHome, you should. It’s cut from the same cloth as Tom’s Hardware Guide and AnandTech but with a focus on server hardware, mostly the low end but they throw in some enterprise stuff, as well.ServeTheHome ran tests comparing the AMD Epyc 7742, which has 64 cores and 128 threads, and the Intel Xeon Platinum 8180M with its 28 cores and 56 threads. The dollars, though, show a real difference. Each Epyc 7742 costs $6,950, while each Xeon Platinum 8180M goes for $13,011. So, two Epyc 7742 processors cost you $13,900, and four Xeon Platinum 8180M processors cost $52,044, four times as much as the AMD chips.To read this article in full, please click here
It’s not just speeds and feeds anymore, it's intelligent software, integrated security and automation that will drive the networks of the future.That about sums up the networking areas that Keerti Melkote, HPE's President, Intelligent Edge, thinks are ripe for innovation in the next few years.He has a broad perspective because his role puts him in charge of the company's networking products, both wired and wireless.Now see how AI can boost data-center availability and efficiency
“On the wired side, we are seeing an evolution in terms of manageability," said Melkote, who founded Aruba, now part of HPE. "I think the last couple of decades of wired networking have been about faster connectivity. How do you go from a 10G to 100G Ethernet inside data centers? That will continue, but the bigger picture that we’re beginning to see is really around automation.” To read this article in full, please click here
We all know that blocking incoming sunlight helps cool buildings and that indoor thermal conditions can be improved with the added shade. More recently, though, scientists have been experimenting with ways to augment that passive cooling by capturing any superfluous, unwanted solar heat and expelling it, preferably into outer space, where it can’t add to global warming.Difficulties in getting that kind of radiative cooling to work are two-fold. First, directing the heat optimally is hard.“Normally, thermal emissions travel in all directions,” says Qiaoqiang Gan, an associate professor of electrical engineering at University at Buffalo, in a news release. The school is working on radiative concepts. That’s bad for heat spill-over and can send the thermal energy where it’s not wanted—like into other buildings.To read this article in full, please click here
HP Enterprise (HPE) has been aggressively promoting its GreenLake IT consumption model since it was introduced last year. GreenLake is a pay-per-use consumption model where the customer does not take ownership of the hardware but merely leases it and pays only for their use, which is metered.Consumption models have become popular among OEMs looking to keep customers that are anxious to get out of owning expensive assets, such as servers. Dell EMC has its own program called Flex on Demand, and Lenovo has ThinkAgile CP.To read this article in full, please click here
Intel announced this week it has begun shipping its 10nm Agilex FPGAs to early-access customers, including Microsoft, featuring the Compute Express Link (CXL), a cache and memory coherent CPUs-to-anything interconnect that has an industry consortium of more than 60 members. The company first announced the chips in April.The Agilex FPGA is the product of the Altera group, which Intel bought in 2015 for $16.7 billion. It sold FPGAs under the Stratix brand name, but this line is the first to come out under Intel ownership. CXL replaces OmniPath Connect, a fabric Intel developed but no one else supported. The company ended support for OmniPath earlier this month in favor of CXL, which has wide industry support.To read this article in full, please click here
Data centers running artificial intelligence (AI) will be significantly more efficient than those operating with hand-edited algorithm schedules, say experts at MIT. The researchers there say they have developed an automated scheduler that speeds cluster jobs by up to 20 or 30 percent, and even faster (2x) in peak periods.The school’s AI job scheduler works on a type of AI called “reinforcement learning” (RL). That’s a trial-and-error-based machine-learning method that modifies scheduling decisions depending on actual workloads in a specific cluster. AI, when done right, could supersede the current state-of-the-art method, which is algorithms. They often must be fine-tuned by humans, introducing inefficiency.To read this article in full, please click here
SAN FRANCISCO – VMware has added new features to its core networking software that will let customers more securely control cloud application traffic running on virtual machines, containers or bare metal. At its VMworld event, the company announced a new version of the company’s NSX networking software with support for the cloud-based advanced load balancer technology it recently acquired from Avi Networks.[ Also see How to plan a software-defined data-center network and Efficient container use requires data-center software networking.]
The load balancer is included in VMware vRealize Network Insight 5.0 and tied to NSX Intelligence software that lets customers optimize network performance and availability in virtual and physical networks. The load balancer includes a web application firewall and analytics features to help customers securely control and manage traffic. To read this article in full, please click here
There are a host of different AI-related solutions for the data center, ranging from add-in cards to dedicated servers, like the Nvidia DGX-2. But a startup called Cerebras Systems has its own server offering that relies on a single massive processor rather than a slew of small ones working in parallel.Cerebras has taken the wraps off its Wafer Scale Engine (WSE), an AI chip that measures 8.46x8.46 inches, making it almost the size of an iPad and more than 50 times larger than a CPU or GPU. A typical CPU or GPU is about the size of a postage stamp.Now see how AI can boost data-center availability and efficiency.
Cerebras won’t sell the chips to ODMs due to the challenges of building and cooling such a massive chip. Instead, it will come as part of a complete server to be installed in data centers, which it says will start shipping in October.To read this article in full, please click here
Big Switch has rolled a variety of new software packages aimed at simplifying customer networks and helping them manage on-premises and hybrid-cloud data-center growth.The products include a version of its Big Cloud Fabric (BCF) for Amazon Web Services virtual private cloud (VPC) management, adding support for Global VPCs (G-VPCs), and a cloud-based version of its Multi-Cloud Director (MCD). VPCs provide the on-demand access to a pool of shared cloud-computing resources.[ Also see How to plan a software-defined data-center network and Efficient container use requires data-center software networking.]
Big Switch’s flagship BCF software lets customers manage physical switches as a single fabric that includes security, automation, orchestration and analytics. BCF can run on a variety of certified switches from Dell EMC, HPE and others. In addition, BCF Controller natively supports integration with various Cloud Management Platforms such as VMware (vSphere, NSX Manager, vSAN) and OpenStack. BCF also supports container orchestrators such as Kubernetes, all via a single interface. To read this article in full, please click here
Data backup and restoration can be somewhat of a black-box effort. You often don’t know whether you fully nailed it until disaster strikes, and there is always room for improvement, especially as cloud and hybrid options grow. We asked four network professionals to share what made them realize they should do more to bolster their organization’s backup and recovery processes, and how they made that happen. Here are their stories.To read this article in full, please click here(Insider Story)
About a cubic yard of freshwater mixed with seawater provides almost two-thirds of a kilowatt-hour of energy. And scientists say a revolutionary new battery chemistry based on that theme could power edge data centers.The idea is to harness power from wastewater treatment plants located along coasts, which happen to be ideal locations for edge data centers and are heavy electricity users.“Places where salty ocean water and freshwater mingle could provide a massive source of renewable power,” writes Rob Jordan in a Stanford University article.To read this article in full, please click here
Extreme Networks has in recent months restyled the company with data-center networking technology acquisitions and upgrades, but now comes the hard part – executing with enterprise customers and effectively competing with the likes of Cisco, VMware, Arista, Juniper, HPE and others.To read this article in full, please click here(Insider Story)
Nvidia is boasting of a breakthrough in conversation natural language processing (NLP) training and inference, enabling more complex interchanges between customers and chatbots with immediate responses.The need for such technology is expected to grow, as digital voice assistants alone are expected to climb from 2.5 billion to 8 billion within the next four years, according to Juniper Research, while Gartner predicts that by 2021, 15% of all customer service interactions will be completely handled by AI, an increase of 400% from 2017.The company said its DGX-2 AI platform trained the BERT-Large AI language model in less than an hour and performed AI inference in 2+ milliseconds, making it possible “for developers to use state-of-the-art language understanding for large-scale applications.”To read this article in full, please click here
Last week, AMD launched the second generation of its Epyc server processor, the Epyc 7002 series a.k.a. “Rome,” and it’s a far cry from the days when it held a release party for the Opteron and no one showed up. These days, AMD has a whole lot of friends.Of course, it helps to deliver a part people want, and it looks like the Epyc 7002 is all that. It builds considerably upon the first generation, code-named “Naples,” delivered two years ago. One chip packs up to 64 cores and two threads per core, double the max of 32 cores in Naples. It has eight memory channels and up to 128 lanes of PCI Express Gen 4.[ Also read: What is quantum computing (and why enterprises should care) ]
The Epyc 7002 achieves this massive core count through “chiplets,” eight small chips in the CPU die with eight cores each and connected by a high-speed interconnect. A single monolithic 64-core die is impractical from a manufacturing standpoint. There is so much more that can go wrong with 64 cores than 16. Plus, AMD is manufacturing this on a 7nm process (Intel is just getting to 10nm), so Continue reading