Get email updates every time we post!
A Server Case Study: HP’s ProLiant ML 350 G4
You’ll find that there are numerous differences between servers (even with entry- level machines) and desktop PCs. The most obvious differences are apparent by examining a server chassis. HP has provided a fourth-generation ProLiant ML 350 to model the server’s innards and to give you a more tangible idea of how the hardware I’ve already discussed works together in an SBS environment. Figure 2-5 shows the front of the server, opened to expose its highly-accessible drive bays.
Check out this post about the HP Proliant ML 350 G4, it’s always great to see what others think about the servers when they’re writing about them.
There has plenty of talk around the green data center and data center energy efficiency, and one of the players has been The Green Grid, a nonprofit group focused on the topic. The Green Grid came out with power usage effectiveness (PUE), which compares total facility power to IT equipment power.
But in the end, what matters is what your data center does, not how much energy it consumes. You might have a PUE of 1.1, but if your servers just sit there idly all day long, who cares? Data centers are built to perform work, and if they don’t do that, energy consumption doesn’t mean squat. Re-enter The Green Grid.
Over the last year or so, there has been discussion over defining data center performance compared to energy consumed, often referred to as a data center’s useful work or data center productivity. The Green Grid has now come out with eight different proposals for “proxies,†which the group describes as approximations for comparing data center production to data center energy consumption. It compares them to the stickers in car lots that claim a certain miles-per-gallon rating, right down to the warning that “your mileage may vary.â€
A value or process to manage data center energy consumption and energy or data center efficiency would certainly make data center management and reporting easier. The challenge represents mainly in the conceptual stage – who determines this value and how do we apply it retrospectively? We need to shift a gear forward, skip to step 2 so to speak. Assign articles in the data center an efficiency standard, and shift the conversation away from the nuts and bolts of the topic to one that’s relevant to the end user. By this I mean, if my DL380 is running my application fine, it works, that it uses 30w more than a newer server might not actually be that big a deal, dependent on who’s paying and the cost of updating or porting the application to a newer platform.
Hopefully we’ll see greater developments in this, but I wonder if we might not do something very simple and therefore inefficient (stick with me). Let’s say I use 500kw, could I divide this by the revenue or value created to give me a per watt efficiency or put another way the amount of power needed to give me my revenue. Therefore, if I replaced the legacy x86 servers with more efficient systems or deployed more virtualization technologies, could this improve my revenue per watt?
Cloud computing is becoming an increasingly hot topic, but companies have been warned to be wary when considering cloud storage services.
The note of caution comes from Tony Reid, UK services director at Hitachi Data Systems, and Ian Osborne, project director of the Grid Computing Now initiative, who both feel that the risks do not yet necessarily outweigh the benefits.
“Cloud computing is still going through the hype cycle, and not a week goes by without yet another company, such as Google, IBM or Symantec, announcing a new cloud offering,” said Reid.
Absolutely, there is still a lot of hype or PR surrounding the cloud concept though this is not necessarily a bad thing. The more we discuss the possibilities, the more we can innovate the platform and create new solutions for the end user community. At the same point as with virtualization (and anything else), we need to be careful that we see cloud as another platform, not an answer to all our problems. Just like grid, web or citrix, moving an application from one platform to another is not going to resolve any inherent issues with the application.
Related to this is a conversation I was having with a journalist the other day “we should just make everything in the cloud”, my response to many a strange look was the following:
Updated on January 26 at 9:50 a.m. PST with additional information about ACP and TDP thermal-envelope ratings.
Advanced Micro Devices on Monday released low-power and high-performance processors that will find their way into servers from Hewlett-Packard, Dell, Sun Microsystems, and Rackable Systems.
The new server processors are updated versions of AMD’s 45-nanometer Shanghai processor.
“When we first came to market, we brought out the standard-power (Shanghai processors) because that’s where the bulk of our market is,” John Fruehe, the director of business development for server and workstation products at AMD, said in an interview. “As always, we follow(ed) up fairly quickly with the HE, which are the energy-efficient models, and the SE, which are the high-performance models.”
The more innovation we realize in the hardware in terms of performance and energy efficiency, the more choice and opportunities for the end user which has to be a good thing for the platform and the community, I’m off to read up more.
In parts 1 and 2 of this series, I discussed two common objections to cloud computing: difficulty of application migration and heightened risk. In this posting, I want to address another common objection to cloud computing, the one that has to do with service-level agreements. I call it:
SLA: MIA
One of the most common concerns regarding cloud computing is the potential for downtime-time the system isn’t available for use. This is a critical issue for line-of-business apps, since every minute of downtime is a minute that some important business function can’t be performed. Key business apps include taking orders, interacting with customers, managing work processes, and so on. Certainly ERP systems would fall into this category, as would vertical applications for many industries; for example, computerized machining software for a manufacturing firm, or software monitoring sensors in industries like oil and gas, power plants, and so on.
Check out this article talking about Cloud Computing, it’s an interesting read. Cloud computing will be an interesting development for the corporate world. How we manage the need to deliver with operating costs, with corporate process and with the power and benefits that cloud could bring could result in a range of new processes, best practice and business opportunities. Certainly outsourcing elements of the infrastructure to a shared cloud environment could be more efficient and possibly reliable than might be achieved internally, how you manage this with any data legislation or issues relating to ownership will depend on your business. I can see though, how for example a cloud platform based on VMware could be very appealing, being able to scale up your server capacity in line with the business need could be a real enabler to your business. I wonder if we’ll see more development of internal clouds, maybe a shared cloud and business specific ones? We’ll see…
BROOMFIELD — Centers that house computer servers and store data aren’t friends of the Earth.
Between 2000 and 2006, data center energy consumption more than doubled in the U.S., according to the Environmental Protection Agency. Consumption exceeded that of all the country’s color televisions.
Sun Microsystems has spent the past 4 1/2 years rethinking the industry.
On Monday, Sun unveiled its latest effort – a data center in Broomfield projected to save the company more than $1 million in electricity costs and 11,000 metric tons of carbon dioxide emissions a year.
An interesting article talking about data center efficiency, in this case talking about Sun’s achievements, do check it out. It’s always good to see what range of technologies, best practice and environmental tools can be used to improve data center efficiency and operating costs.
BARCELONA, Spain (Cisco Networkers)- January 26, 2009 – Cisco today announced it has collaborated with NYSE Technologies to develop an Ethernet-based integrated messaging solution for the financial services industry that accelerates automated trading environments and improves end-to-end application performance using industry standards-based technologies. The solution supports the Cisco® Data Center 3.0 vision to help its customers build next generation data centers that support their business goals. The solution tested by Cisco combines NYSE Technologies’ Data Fabric high performance middleware, NetEffect Remote Direct Memory Access (RDMA)-capable 10GbE Network Interface Cards (NICs) from Intel®, Cisco Nexus 5000 Series switches and Cisco RDMA Accelerated Buffers (RAB) drivers to accelerate market data infrastructures without requiring customers to implement new transport technologies and without having to change their applications.
10GB Ethernet based solutions bring new opportunities for trading platforms and for virtualization/consolidation, this article talks about a messaging solution for the financial services sector, do check it out.
HP is releasing its first-ever line of devices designed specificially to make data center connectivity for easier and more efficient.
Hewlett-Packard ProCurve is extended its adaptive network mantra into the data center with its first-ever line of switching platforms designed for the data center environment.
“ProCurve is certainly no stranger to the data center environment – it has been deployed in data centers in the past,†said Matt Zanner, worldwide director of Data Center Networking for ProCurve. “This is the first attempt at getting our hands on products to optimize data center capabilities.â€
The 6600 series consists of five top-of-rack server edge switches featuring the same firmware as the ProCurve 5400 and 8200 series switches to enable seamless connectivity. “These are leveraging same underlying architecture as rest of Intelligent Edge products,†Zanner said.
In addition, the switches feature front to back, reversible airflow that helps keep them cooler and reduces energy consumption.
ProCurve also is rolling out the Data Center Connection Manager, a management tool that reduces complexity at the management layer, according to the company.
“This allows for network connections to be applied in an automated and easy-to-manage fashion,†Zanner noted. “Administration teams, both server and network, will now be able to work together in a seamless fashion. It’s about furthering business operations.â€
Anything we can do to make systems deployment and administration has to be a good thing, the more innovation we have in this area, the more we can centrally manage and support our systems, to reduce our operating costs and improve our ability to resolve an issue and maintain service. Do check it out.
Rackable Systems Inc. says that its new 12-volt servers aim to turn virtualization on its head, but this hardware-oriented approach might be a hard sell for some data center managers.
The single-socket physical servers slide into Rackable’s CloudRack enclosure trays and run divided workloads. They are based on what the company calls its “MicroSlice” architecture and are Rackable’s smallest servers to date. Rackable is known for half-depth servers that take up less data center space than traditional rack servers.
Rackable positions the new machines as an alternative to software-based server virtualization. While virtualization software consolidates servers to run applications on independent virtual machines (VMs) that reside on fewer physical servers, the MicroSlice servers are a bunch of single-socket nodes that run divided workloads in the CloudRack enclosure. Rackable has termed this hardware approach physicalization.
This sounds great, I’ll need to read up more, these single socket servers might be an ideal solution for specific customers with specific project needs or energy efficiency. I’ll need to check it out.
January 26, 2009 (Computerworld) How did Citigroup‘s green IT initiative begin? I began working on this in April or May [of 2007. Former Citigroup CIO] Marv Adams had a strong commitment [and] asked me to look into this. I ended up speaking to people from around the bank and got the program approved by the leadership council in September 2007. The bank has committed $50 billion toward sustainable efforts over the next 10 years. We also have a plan to reduce our own greenhouse gas [emissions] by 10% by the year 2011. And we’re operating from a 2005 base.
One of our strengths is our global approach, to see the big picture and see the points of interconnectedness to make sure everyone is talking.
An interesting article talking with the former CIO of Citigroup, it’s always good to see what different organizations, CIO’s see as their big issues, what their drivers are in terms of cost and the environmental factors – the green data center etc, do check it out.
