IDEAS has released an update to the Server Efficiency Trend that shows that long-term server efficiency, defined as server performance per watt, continues to improve, although the growth rate may be slowing.
Figure 1. IDEAS Server Efficiency Trend (Server Performance/Watt)
Continue reading "IDEAS Trend Analysis Shows Server Efficiency Growth Rate May Be Slowing" »
Hitachi Data Systems (HDS) recently introduced the Hitachi Unified Storage (HUS) 100 platform to replace its Adaptable Module Storage (AMS) 2000 family of midrange products. HUS employs the gateway-based architecture of HDS's NAS solutions, featuring an enhanced version of the AMS 2000 FC/iSCSI SAN arrays attached to the Hitachi NAS Platform (HNAS) 3080/3090 gateways. The key enabler of this unified storage solution is the Hitachi Device Manager software, which integrates more deeply with HNAS in order to provide converged management for file and block storage. While HUS already offers a converged software-licensing scheme for block- and file-based features, the updated Device Manager software will now technically enable the same level of unified manageability for all HNAS-based solutions, benefitting many existing customer installations.
Architecture of the HUS 150 (source: HDS)
Continue reading "HDS Device Manager Unifies Block and File Management" »
IDEAS has released an update to the Server Performance Trend, which predicts that server performance will double from today’s levels during the second half of 2013. The Server Performance Trend is powered by RPE2, the atomic unit of compute, and provides a “whole market” view of how server performance is trending.
Continue reading "IDEAS Research Predicts Server Performance Will Double by the End of 2013" »
Emulex has announced a new line of 10 Gigabit Ethernet (10GbE) network adapters called OneConnect Network Xceleration (NX). The technology comes from Myricom, a networking vendor focused on high-performance computing (HPC) and other niche markets. By combining Emulex’s position in the mainstream network interface card (NIC) market with Myricom’s low-latency technologies, the partnership helps build credibility for 10GbE in specialized markets like high-frequency trading (HFT).
Continue reading "Emulex-Myricom Partnership Makes Low-Latency Ethernet Mainstream" »
Over the past few years, HP has been the “package king” of proactive maintenance, leading the way with transparent, part-number-based, fixed-specification, fixed-price, proactive support offerings. The company has built a portfolio of such support services targeted at large and small installations. The latest addition is Proactive Care, which emphasizes automation and remote support, and which allows HP to target volume sales through simplicity in packaging and pricing.
What does Proactive Care mean for HP, and which customers stand to benefit from it?
Figure 1 – Entry-Level, Three-Year Proactive Support Costs for a DL380p Gen8 Server
Continue reading "HP Lowers the Bar to Entry for Proactive Services" »
Several IDEAS analysts attended Oracle Industry Analyst World 2012 at its campus in Redwood City, CA a few weeks ago. At this annual event, Oracle provided an update on its strategy to maximize the value that can be gained from integrating its hardware and software assets into complete solutions. The event focused primarily on Oracle’s customer momentum, as well as a review of some hardware and software announcements made since last year’s analyst conference. Oracle took this opportunity to remind analysts that its strategy feels solid and that it does not need to be tweaked on a yearly basis.
The USA 76 America's Cup Class Racing Yacht Used by Oracle Racing
Continue reading "Oracle Systems Strategy Remains Steady, Focus Turns to Execution" »
It is seductive to say that "more is better," but when it comes to datacenter infrastructure, the number of servers is a misleading metric to judge value. Public cloud computing is sold in nebulous units called "instances" that often come without capacity or throughput specifications. This makes it hard for users to do apples-to-apples comparisons. In order to compare the infrastructure as a service (IaaS) component of datacenter as a service (DCaaS), one needs to measure compute capacity in terms of performance metrics and not "number of instances" or "number of servers." Amazon's standard metric for compute instances, elastic compute units, shows the different compute capacity one gets for a given budget. But an industry-wide standard performance metric is needed for a complete understanding of cost and capacity.
Figure 1. Buying $100 of Amazon EC2 Cloud: Number of Instances vs. Aggregate Compute Power
Continue reading "Buying Datacenter as a Service? Measure Capacity, Not Units" »
After more than three years of development, $2 billion of investment, and its acquisition of Blade Network Technologies, IBM has jumped headfirst into the market for preintegrated systems, which has been dominated until now by Cisco, HP, and Oracle. In a recent series of high-profile worldwide announcements, IBM revealed the PureSystems family of systems, which combines processors, storage, networking, and application software within a single, integrated form factor. Rather than labeling its new offerings as "converged systems," which has been the term associated with existing products from competitors, IBM has defined a new product category for PureSystems that it calls "Expert Integrated Systems." IBM describes this new category as providing built-in expertise (that is, embedded best practices for deploying specific workloads); a new design that fundamentally targets delivery of systems in integrated form; and a simplified experience throughout the lifecycle of applications.
Continue reading "IBM PureSystems Raise the Bar for Preintegrated Systems" »
