Adam_a

Mod
  • Content Count

    2665
  • Joined

  • Last visited

  • Days Won

    15

Everything posted by Adam_a

  1. To address this, AWS released the new G4 instances that leverage the latest generation NVIDIA T4 GPUs, custom 2nd Generation Intel Xeon Scalable (Cascade Lake) processors, up to 100 Gbps of networking throughput, and up to 1.8TB of local NVMe storage, to deliver the most cost-effective GPU instances for machine learning inference. The new G4 instances can deliver up to 65 TFLOPs of mixed-precision performance, great for inferencing. G4’s can also be used on cost-effectively for small-scale and entry-level machine learning training jobs that are less sensitive to time-to-train. For graphics-intensive workloads, the new instances offer up to 1.8x increase in graphics performance and up to 2x video transcoding capability over the previous generation G3 instances. AWS Releases EC2 G4 Instances With NVIDIA T4 Tensor Core GPUs
  2. The AMD EPYC 7002 came out with the promise of more performance with fewer sockets. Now that VMware fully supports the new CPUs, we continue with our testing to show the promise of the new processors. For this review we were able to test three distinct AMD EPYC CPUs in a traditional VMware environment. This testing wasn’t to show which is better, that is typically the one with the highest core count. Instead we tested the different CPUs for readers to get an idea what to expect, so they can better choose the processor that is right for their use case. AMD EPYC Rome VMware Review (7702P, 7402P, and 7302P)
  3. The new Intel Stratix 10 DX FPGAs are the first, and so far only, FPGAs designed to support Intel Ultra Path Interconnect (Intel UPI). Intel is claiming that when paired with future Intel Xeon processors that also support UPI, the new cards will have 37% lower latency compared to cards that can only communicate over PCIe-4. Like all Statix 10 products, the DX cards are built on a 14nm tri-gate architecture. Intel Ships Early Access Stratix 10DX FPGAs
  4. Most of the improvements in this OpenStack Platform 15 look to be security related, although several straddle the line between security and performance. OpenStack 15 does provide support for NVIDIA GPU/vGPUs to support the growing market demand for artificial intelligence (AI) and machine learning (ML) workloads. One of the bigger performance wins comes through improved crypto offload for IPSec VPN to NICs, allowing companies to implement robust security while retaining better network performance. Another improvement which straddles the line between a security and performance improvement is the enhanced TCP stack which should improve application uptime as well as workload security. Another source of potential performance gain is increased backend support for dedicated hardware cryptographic solutions offering more secure messaging while freeing up more flexible system resources. Red Hat Releases OpenStack Platform 15
  5. Kubernetes is an open-source container-orchestration system and has been widely popular since it came out. As more cloud native applications come out, Kubernetes becomes more appealing. And now almost every major vendor has some form of support of the system. Kubernetes 1.16 Released
  6. Synology started nearly 20 years ago with its core mission being to provide data management and security tools for an ever-evolving technology landscape. That company feels confident that it has stuck to this mission over the years while evolving along the way with an industry that is almost always in flux. Synology went on to prove this point by showcasing its next generation files server technology and a demonstration of its license-free protection suite Active Backup. The company spoke about its public cloud service, C2, and its hybrid cloud model for data management and Synology’s cloud data protection service, C2 Backup. Synology 2020 NYC Sees Several Announcements
  7. Gigabyte has a strong partnership with AMD and is a leading systems vendor for the EPYC 7000 Series platforms. With these six new systems being the second wave of lineup released, the GIGABYTE EPYC 7002 portfolio totals nearly 30 systems and includes rack mount servers, GPU servers, and high density multi-node servers. Offering a large and diverse portfolio GIGABYTE looks to be a top vendor in the AMD EPYC 7002 Series system market. GIGABYTE Expands AMD EPYC Rome Server Systems
  8. The nanoEDGE is a small form factor SDDC appliance that runs the full VMware vSphere, VMware vSAN, and VMware NSX in a low cost, small physical volume appliance. The nanoEDGE provides enough compute to run 25-50 VMs. The nanoEDGE is described as 2.5 servers: 2 servers provide vSAN and NSX and the third (0.5) server provides the vSAN Witness Appliance. This allows the solution to utilize 10GbE back to back connectivity for vSAN and vMotion traffic while keeping costs in check. VMware nanoEDGE A Small, Low Cost SDDC
  9. Together the companies in their joint effort provide cost-effective backup and recovery for both on-premises and cloud environments. Veritas NetBackup provides snapshot-based data protection. Integrating NetBackup into Pure Storage’s FlashArray has helped customers guard against ransomware and data loss. Pure Storage integration of FlashBlade into Veritas enables data protection and rapid restore for critical workloads. Additionally, customers can leverage FlashBlade as a Data Hub for streaming analytics and AI cluster projects with Veritas NetBackup providing data protection. Veritas & Pure Storage Collaboration Expansion
  10. As we said at the launch of the new CPUs, the architecture of the second-generation AMD EPYC is all about performance. It is the world’s first high-performance x86 7nm server CPU. AMD has claimed to improve the CPU core execution engine over the original “Zen” core. Additionally, EPYC Rome has twice the density, 1.25x the frequency (at the same power), and uses half the power (at the same performance of the previous model). AMD goes on to state that there is more on-chip bandwidth in all compute modes. The new CPUs also have high memory speed and bandwidth with a maximum theoretical memory bandwidth of 410GB/s (45% higher than Intel Xeon 2nd gen Scalable CPUs). AMD Has A Rome In Rome Day Focusing On EPYC 7002
  11. Oracle’s big thing over the last few years has been autonomous. Chiefly its Autonomous Database (that has seen very good worldwide adoption) but also Oracle Autonomous Transaction Processing and Oracle Autonomous Data Warehouse. For its Autonomous Database, Oracle is expanding the capabilities once more. The company is now offering a dedicated deployment option for Oracle Autonomous Database. This option will give customers high levels of security, reliability, and operational control regardless of the class of database workload. Oracle Announces New Innovations At OpenWorld
  12. Vertica analytics platform in Eon Mode for Pure Storage is a mouthful of words, so let's pull it apart and see what each piece means. Vertica is a set of tools to analyze information stored in an SQL database. Eon Mode is a set of configurations for this set of tools. What makes Eon Mode special is that it allows Vertica customers to separate compute resources from data storage and better control costs. Previously this mode, and thus the cost savings were focused on the AWS ecosystem, but now Micro Focus is extending the same cost savings opportunities to Pure Storage. Micro Focus Releases Vertica in Eon Mode for Pure Storage
  13. The partnership allows both companies to reach across the aisle and take advantage of each other’s technology. For instance: Oracle becomes a partner in the VMware Cloud Provider Program and Oracle Cloud VMware Solution will be sold by Oracle and its partners. This Oracle Cloud solution is built on VMware Cloud Foundation and is said to deliver a full stack SDDC that includes VMware vSphere, NSX, and vSAN. Now customers should be able to seamlessly migrate workloads between on-premise environments and Oracle Cloud. VMware Cloud Foundation On Oracle Cloud Infrastructure Announced
  14. Of the five new PowerEdge Servers announced, only the two single-socket servers are currently available. The R6515 is a 1U rack server with a single socket. The R7515 is a similar design, scaled up to a 2U form factor. Dell EMC expects the R6525 and C6525 servers next month. The PowerEdge R6525 is a 1U rack server with two CPU sockets. We have very little solid information on the C6525, but we do know it also has two CPU sockets. It appears to be a 2U form factor. The only other c-series PowerEdge server currently available is the c4140, a 1U form factor rack server so it is difficult to make intelligent guesses as to its exact specs, but it is probably safe to anticipate a design that supports a relatively large number of GPUs or other ancillary processing cards like FPGAs. The fifth and final server also has the vaguest availability date, being projected out to sometime next year. The R7525 will be a two CPU rack server in a 2U form factor. Dell Releases EPYC Rome PowerEdge Servers
  15. EDGEMATRIX will initially focus on the Japanese market, where it has already raised $9 million in Series A funding. Cloudian management is betting heavily on the company’s success with both Cloudian CEO and co-founder Michael Tso and Cloudian board director Jonathan Epstein investing in it. The new company plans to leverage Cloudian’s HyperStore platformto manage large data sets. EDGEMATRIXwill also be inheriting the work Hiroshi Ohta, Cloudian co-founder, and Cloudian teams have done over the past two years to provide AI solutions for such use cases as: targeted advertising using video analysis of the make, model and year of cars on a highway to display targeted ads on a nearby billboard; quality inspection by identifying and removing products that failed to get wrapped in an automated assembly line process; and automated traffic congestion management by counting cars on a road as a basis for changing turn lanes to improve traffic flow. Cloudian Launches Edge AI Subsidiary
  16. Oracle Exadata X8M is a combination of industry-standard Intel 2nd generation Xeon Scalable processors, Intel Optane DC persistent memory, and 100 gigabit RoCE networking. All of this performance-based hardware is designed to support demanding Online Transaction Processing (OLTP), analytics, and mixed workload database requirements, as well as database consolidation and in-database machine learning. Several activities involved in the above are extremely sensitive to latency including high-frequency stock trading, IoT data processing, real-time fraud and intrusion detection, financial trading and applications requiring real-time human interactions. With the integration of Optane, users should see latency improvements in all of the above. Oracle Integrates Intel Optane Into Its Exadata X8M
  17. The release of the ThinkStation P920 Rack offers professional workstation performance in a fully scalable and secure 2U form factor. Leveraging low latency remote access technologies, users are given true freedom by being able to access extreme computing power anywhere and collaborate uninterrupted. This ThinkStation joins their portfolio of P series workstations and is the first rack-optimized form factor in the series. The P920 Rack offers a dual processor platform with Intel Xeon Scalable processors, NVIDIA enterprise grade graphics, up to 3TB of ECC memory, and up to 112TB of enterprise storage. Additionally, enjoy 99.99% availability as it is backed by hallmark ThinkStation brand reliability with hot swappable and redundant features and datacenter-class security. Lenovo ThinkStation P920 Rack Announced
  18. Disney and Microsoft are both industry leaders in their respective fields. Disney brings to the table a history and passion for innovation as well as industry influence in defining new ways to create content in magnitude. Microsoft brings to the table it’s Azure Cloud Platform that features hyperscale capacity, global distribution, industry-leading storage and networking capabilities. Partnering with Disney unlocks a new opportunity for growth in the media industry for Microsoft. Walt Disney & Microsoft Azure Announce Partnership
  19. The Western Digital Ultrastar DC SN630 is the new data center NVMe SSD that comes in two flavors: read-centric and mixed-use. The drives come in capacity ranges from 800GB-6.4TB for the mixed-use and 960GB-7.68TB for the read-centric. The drive leverages Western Digital’s controller, firmware, and 64-Layer BiCS 3D NAND. All drives offer ISE, which is great for redeployment or retirement. Another security feature is the use of secure firmware downloads with RSA authentication to ensure the SN630 runs authentic firmware only. Since the SN630 is vSAN certified, we tested it within the context of VMware vSAN to see how it performed. Western Digital Ultrastar DC SN630 VMware vSAN Review
  20. With today’s newest solution added to their portfolio Supermicro now has nearly 20 unique SuperWorkstations with a range of specifications. Systems have either single- or dual-processor platforms, robust memory capacity – up to 4 DDR4-2933MHz memory in 12 DIMM slots, 7 PCI-E 3.0 slots for GPU/Coprocessor, dual Gigabit Ethernet LAN port, and single 10 Gigabit LAN port. The takeaway from the solution announced today, the SYS-5049A-T, is that it is a single processor platform with a high core count, significant memory capacity, multi-GPU, and supports ultra-fast NVMe storage. Supermicro Launches SYS-5049A-T SuperWorkstation
  21. Located in Sunnyvale, California, Rafay is led by CEO Haseeb Budhani and VP of Engineering Hemanth Kavurulu. Both men cofounded the company after working together at Soha Systems on systems development, cloud networking, and network security. Soha was acquired by Akamai Technologies in 2016. Rafay’s main focus as a company is the above mention SaaS for container lifecycle management. Rafay Launches SaaS Automation Framework For Containerized App Lifecycle Management
  22. Designed to match the mission-critical capabilities of IBM Z and LinuxONE servers, the IBM DS8900F family is built with the most advanced POWER9 processors. The DS8900F is the result of deep collaboration between the IBM Storage and IBM Z teams to deliver unique business value for mainframe deployments. The company states that the new storage systems are able to offer transparent hybrid multicloud connectivity for infrastructure simplification, massive capacity, and data protection through the cloud. The new family will replace the DS8880 line. The IBM DS8900F will off a DS8910F model for entry level and the DS8950F that will consolidate all mission-critical storage workloads for IBM Z, IBM LinuxONE, IBM Power Systems, and distributed environments under a single all-flash storage solution. IBM Announces DS8900F Storage Systems & z15
  23. Customers that leverage AWS Outposts can local launch a range of Amazon EC instances (including C5, M5, R5, I3en and G4, both with and without local storage options) as well as Amazon EBS volumes. Amazon plans to have many other services ready at general availability, such as Amazon ECS and Amazon EKS clusters for container-based applications, Amazon EMR clusters for data analytics, and Amazon RDS instances for relational database services with more planned in the future. Outpost works as an extension of the AWS Region of the user’s data center. Services running on Outpost can seamlessly work with any AWS service or resource running in the cloud. New services launched by AWS will also work with Outposts. The company also states that API calls will be logged via CloudTrail automatically and existing CloudFormation templates will work as well. More Details On AWS Outposts
  24. It’s been some time since we’ve been this deep with anything in the Microsoft storage-centric stack; and boy, are we glad to be back. With the rebranded Microsoft Azure Stack HCI solution, Microsoft has done something that’s so basic and fundamental, it’s easy to under appreciate. Microsoft has made their HCI solution dead simple to operate without overlaying anything to dampen performance. As seen in our numbers, the DataON cluster we’ve been testing has posted tremendous numbers, the fastest we’ve seen in a mid-market 4-node HCI cluster. To be fair, we’re not even testing the latest and greatest hardware from DataON either. While this config is clearly no slouch, complete with Intel Optane DC SSDs, DataON offers faster solutions that take advantage of Intel Xeon 2nd Generation CPUs, persistent memory and faster networking. The fact that there’s even more performance available in an Azure Stack HCI solution is exciting, but it’s also important to remember the solution can scale down as well to deployments as small as two node HCI that can be configured switchless for a low-cost edge or SMB solution. Microsoft Azure Stack HCI Review (DataON HCI-224 with Intel Optane NVMe)
  25. Data is growing and becoming more valuable. Companies like Spectra Logic have been refining the ability to store massive amounts of data, however, most data stored is cool or inactive. Up to 80% of data that is gathered my not be of immediate use but is still valuable. Storing this cool data on primary storage is expensive and not very useful in the long run. Spectra StorCycle software automatically tiers the inactive data to a more cost-effective storage medium. Spectra Logic StorCycle Unveiled