Home > Intel Communities > Open Port IT Community > The Server Room > Blog > 2008 > February
0

 

Virtualization is without a doubt a very hot topic these days. Companies continue to look to server virtualization to increase the utilization rates of their systems and lower overall deployment and management costs. The basic model of a virtualized server is depicted below:

 

 

 

 

 

 

Essentially, you have a VMM (Virtual Machine Monitor) SW layer that talks between hardware and software and allows each virtual machine to successfully use what it thinks is one network port. This is a pretty straightforward model and it directly addresses the general reason for virtualization which is that generally the server may not be utilizing its processing power in full and is thus wasting CPU cycles.

 

 

There is an interesting result of this consolidation onto a single physical box with several Virtual Machines. In addition to consolidating CPU processes, you also effectively consolidate I/O bandwidth and switch processing capabilities onto the same platform. The overhead of this switching limits your bandwidth, adds CPU overhead, and effectively reduces the benefits of server virtualization. In some cases you may have a new problem in having created an I/O bottleneck.

 

 

This makes a lot of sense if you think about the fact that in essence, what you are doing is merging 5-10 machines that each had 1 or 2 ports of Gigabit Ethernet (all connected via a switch) into a single machine. This new server probably needs to have at least 6 ports or more of Gigabit Ethernet and may even require 10 Gigabit connections just to be able to support the new consolidated workload.

 

 

Enter Virtual Machine Device Queues (VMDq):

 

 

In order to help the I/O congestion associated with the additional VMM software switching in a virtualized environment, Intel implemented a technology called VMDq in our latest Ethernet NICs and silicon. VMDq is a technology specifically designed to offload some of the switching that was done in the VMM to networking hardware specifically designed for this function. This drastically reduces the overhead associated with I/O switching in the VMM which greatly improves throughput and overall system performance.

 

 

Below is a diagram that summarizes the new virtualized server stack with VMDq enabled:

 

 

 

 

 

 

On the receive path, VMDq provides a hardware ‘sorter' or classifier that essentially does the pre-work for the VMM of directing which end VM the packets should go to. The NIC or LAN silicon is performing a hardware assist for the VMM layer.

 

 

On the transmit side, the packets are serviced round robin style to avoid "head of line" blocking and alleviate potential quality of service (QoS) issues.

 

 

The immediate question I expect is "So, don't the VMM vendors have to support this?" And the answer is yes. Intel is supporting this feature today on shipping platforms, but you do need to work closely with the VMM vendor to make sure the whole stack works as designed.

 

 

Just this week Intel announced that our VMDq capability will be supported in VMware's upcoming ESX release. This is certainly a big step towards wide support of network virtualization performance enhancing features.

 

 

Ethernet technology has grown and become more important over the last 25 years, and the trend appears to be continuing on course.

 

 

Ben Hacker

 

 

--

 

 

For more details on VMDq, there is a VMDq Whitepaper, and an Intel® VT for Connectivity Datasheet located on our website.

 

 

0 Comments 0 References Permalink
3

During my blog post last week titled "[processors-cpus-cores-arent-they-all-the-same|p-10936]", I got a request to talk more about server terminology, going beyond the CPU. So here we go.

 

The processor does all your computational calculations. The chipset is attached to the processor and manages information and data flow from the processor to the other system components like memory and other I/O (input / output) devices. If the processor is the "brain", then the chipset is the "heart". Memory and I/O devices are like "arms" and "legs" - you need them all working together. The term chipset, memory controller, and memory controller hub (MCH) are often used interchangeably

 

At Intel, we refer to the platform as the combination of all of these devices that comprise a server. From a hardware perspective, I see the following terms used interchangeably: platform, solution, system, server, workstation, and there are probably others I'm missing. However, I have also seen platform refer to the software stack, as well as the complete hardware and software solution together.

 

Another term that you may also here is form factor. Form factor refers to the size and shape of the final system you would buy from a manufacturer. The most common server form factors are rack mount, pedestal, towers and blades.

 

 

3 Comments 0 References Permalink
0

Maybe it's the free drinks of the welcome reception but 2 hours after the breakout sessions have finished the show floor is still buzzing with deep and serious conversations in every corner.

 

 

Green is the overarching theme of this years EMEA VMworld in Cannes, from the issues of data centre power all the way to the powerpoint template used for every presentation thru to the shirts worn by VMware staff. General messages from Diane Green's ( there's that green element again ) opening keynote were around - flexible & dynamic IT, Disaster Recovery, High Availability, Streamlined software development and client virtualisation. BT, IBM, HP & Dell provided walk-on parts during Diane's keynote with BT talking to how they are implementing a fully virtualised infrastructure across their data centres.

 

 

Walking the show floor key major themes are management, data backup & security, Virtual desktop and of course Green IT with everyone having their own spin on what this means. At previous VMware events in EMEA around 50% of the exhibitors have been focused on Virtual desktop, this year it feels that there is more of a bias towards management and backup, but this may just be as a result of VMware expanding their focus beyond their previous technical audience in EMEA. Either way VDI is still a key element of many of the sessions and the show floor booths. Its difficult however to judge how many end-customers are really taking this path with their client infrastructure - I'd be interested to here your thoughts on whether this really is a viable usage model for client systems.

0 Comments 1 References Permalink
6

I get questioned often about the difference between these terms and it can be confusing. Now that we are in the era of multi-core, let's explore common terminology. What is the difference between a processor, CPU, a chip, a core and a socket? And how is threading different?

 

Processor / CPU: This is what Intel makes and OEMs design into their systems.
Processors and CPUs are sometimes referred to as CHIPs
*Sockets:* The physical location on the system board where the processor/CPU goes. Sockets are increasingly used to describe a servers capability. A 4S (4 socket) server supports up to 4 CPUs inside. Sometimes this might also described as to as 4w (wayness) or 4P (processor) server.

Cores: The number physical processing units contained within the processor. There can be one, two, four or more ...
*Threads:* Some Intel processors support multi-threading technology. This is simply the ability to run more than one software thread on a core (Single threaded means one stream of software per core at a time) (Multi-threaded means more than one stream of software is executed in parallel)

So ... Processor, CPU, Socket, and Chips are terms that are often used interchangeably. Cores and Threads are both features inside the processor. Was this helpful to you ? Let me know. Chris

6 Comments 0 References Permalink
0

While Intel is certainly most widely known for manufacturing our extremely complicated CPUs that are the brain of many computing platforms worldwide; there are several other products and technologies that people at Intel have been involved in for many years which are critical to computing environments everywhere. As a person who has been working in various networking and manageability roles at Intel since 2001, I'd like to take a little time to focus on Intel's history in the Ethernet market since its inception more than 25 years ago and focus a little on where the market might be going in the future.

 

Below is an image that tries to capture the key highlights of Intel's specific involvement in the Ethernet market over the last 3 decades:

 

 

 

As you can see the Ethernet market has come along way from clunky multi-chip 10Mpbs solutions from more than 25 years ago all the way to Quad Port Gigabit and Dual Port 10 Gigabit designs that are prevalent today.

 

Moving into the future the Ethernet market is growing increasingly more complicated by the year with new capabilities and features targeted specifically to support server virtualization, infrastructure convergence, enhanced storage technologies, and the continued importance of power efficiency of the overall compute infrastructure. Each of these innovations and changes will have a big impact on the overall structure and design what servers and datacenters will look like in the future. My colleague Ken Lloyd gave his thoughts on how 10 Gigabit technologies will provide I/O convergence and overall cost savings for computer networks in the future and there are clearly lots of interesting things going on right now.

 

Over the next several months I plan to try to go more in depth on many of the exciting developments taking place in the Ethernet market and to hopefully shed some light on some of the changes that are coming our way.

Stay tuned in the coming weeks!

- Ben

0 Comments 0 References Permalink
3

 

Datacenter Power Management: Power Consumption Trend

 

 

Jackson He

 

 

As the internet services grow and the more users embracing internet - approaching 1 billion connected users, one of the biggest challenges for data-center operators today is the increasing cost of power and cooling as a portion of the total cost of operations. As shown in Figure 1, over the past decade, the cost of power and cooling has increased 400%, and these costs are expected to continue to rise. In some cases, power costs account for 40-50% of the total data-center operation budget. To make matters worse, there is still a need to deploy more servers to support new business solutions. Data centers are therefore faced with the twin problem of how to deploy new services in the face of rising power and cooling costs. In a recent survey of data centers 59% identify power and cooling as the key factors limiting server deployment.

 

 

 

 

Figure 1: IDC Report of data center cost structure and trend

 

 

At the same time with the increased energy cost and awareness of global warming, there is increased regulatory scrutiny around both idle and max power of servers and clients (desktops and laptops). The "green awareness" datacenter is no longer a "nice to have" feature, but of necessity of business operation and environmental regulatory compliance. Figure 2 highlight the world-wide existing and emerging regulations on power and energy consumption. Future datacenters have to be able to clearly measure and proof regulation conformance in order to operate properly.

 

 

 

 

Figure 2: Existing and emerging energy and power regulations

 

 

To sum it up, the power management trends for future datacenters are multifaceted and will not be covered by a single company or a single business segment. They could be summarized in the following areas:

 

  • At environment level: conform to increased government regulations on energy and power and increased power constraint (limited available power) - need innovative ways to conform "green datacenter" regulations, while deliver great values to business.

  • At the datacenter level: more computing power is needed with increased demand; emergence of mega datacenter and modular datacenter (datacenter in a container); the overall power and cooling distributions need to match the increased need - new datacenter designs and power/cooling management needed.

  • At rack level: higher power density and higher server density per rack is needed to pack more computing power for a given space and cooling; workload balance between racks to increase power efficiency and overall datacenter reliability - need effective rack-level power and cooling monitoring and dynamic management capabilities

  • At server level: need lower idle and max processing power, so that platform power consumption trend is more linear with platform performance; dynamically adjust power consumption based on policy and workload - need more server-level instrumentations for power/cooling monitoring and more control knobs to dynamically optimize power and performance.

 

I hope you agree with me of the overall datacenter power management trends at datacenters in the coming year. These trends pose challenges for each of the areas listed above. These challenges also mean opportunities for innovative solutions to thrive. I'd like to listen to your feedback about these trends. I will talk more about challenges and potential solutions in the upcoming blogs. You are welcome to share your thought of where you believe the datacenter power management is going. Thanks a lot.

 

 

3 Comments 0 References Permalink
3

 

Here's the 3rd follow-up post in my 10 Habits of Great Server Performance Tuners series. This one focuses on the third habit: Know Your Platform.

 

 

 

As we learned in my last blog, we should start our server performance tuning by looking for system-level bottlenecks. This involves understanding exactly how data flows into and out of your platform - and to do this, you need a block diagram. A block diagram shows the major components on the server's motherboard and the paths between them. From a good block diagram you can derive the maximum data transfer rate (aka bandwidth or throughput) achievable as data flows along those paths.

 

 

I usually look at my block diagram before beginning system tuning in order to identify potential bottlenecks. But some people use them in parallel: they measure the bandwidth of various parts of the system and then confirm what they see using the block diagram. You can determine if various parts of your system are heavily stressed, bottlenecked, or lightly utilized. In general you want to trace the path from where data enters your server (NIC, HBA, etc) up to the processor and back to memory or out of the server. The paths connecting one component to another are commonly known as buses. For each bus, multiply the speed by the width to determine the maximum potential bandwidth.

 

 

Let's use the block diagram for the Intel S5400SF server board as an example. It has 2 FSBs, each capable of 1333 or 1600 Mega-Transfers/second (MT/s). Each transfer on the FSB is 64 bits (8 bytes), so 8 bytes * 1,600,000,000 transfers gives a maximum theoretical bandwidth of 12.8GB/s per FSB segment. Keep in mind though that in reality a bus will not achieve its theoretical maximum bandwidth - depending on the type of bus it will probably realize 66-80% of the possible throughput.

 

 

 

 

So, where do you find these diagrams? If you are using an Intel server platform, the block diagrams can usually be found in the technical product specification for each board. If you purchase a platform from one of our OEM partners, ask your salesperson where to get it.

 

 

Look at the maximum bandwidth achievable on each link your data will travel over to gain a deeper understanding of how your workload will run on your platform.

 

 

Keep watching The Server Room for information on the other 7 habits in the coming weeks.

 

 

 

3 Comments 0 References Permalink
0

Data Center Fabric

Posted by Ken Lloyd Feb 13, 2008

Whenever I see the word fabric, I immediately think of cloth, which immediately takes my mind to "helping" pick out curtains, which puts me in immanent boredom mode. Pardon the offense to those of you that actually know what color curtains you own, but there are some very cool things happening in data center communication.

 

Today the boxes in the data center (servers, storage, switches, ...) communicate over some combination of medium and protocols. While some protocols have become less common, (DECnet, Token Ring) there is still a bunch of InfiniBand, Fiber Channel, and Ethernet. Guess what, Ethernet is going to win. Ok, that was an unsupported prognostication,,, but Ethernet has won in every other arena it has entered.

 

Ethernet is what I really want to talk about today. There are a series of changes happening that allow Ethernet to be cast in the role of data center fabric. The first is simple throughput - 10GB. 10GB has the capacity to support the needs of detached storage. The thing that really makes this possible is a QOS feature called "priority pause". This extension to the Ethernet standard enables Ethernet to support QOS for differentiated services and to minimize or eliminate packet drops.

 

This new "Ethernet" enables rich SOE ( Storage over Ethernet) beyond iSCSI to FCOE( fiber channel over Ethernet). Intel has open[-sourced|http://open-fcoe.org/] FCOE software, and the network community is actively discussing the future of Fiber Channel.

 

 

Consolidating on 10gb reduces required port counts, and a single protocol reduces server hardware and switch infrastructure. All of this saves energy and simplifies data center wire management. These are good things. The extensions to the Ethernet specs were the result of collaboration between Intel and other industry leaders. This new spec should make it simpler to choose which curtains will go best in the data center

 

 

Intel is the leader in the add on server NIC business has great products available in the 10gb NIC space. The Intel NICS, when used on and Intel based platform, also support VMDQ - part of Intel's "Virtualization Technology for Connectivity". At VMworld Intel demonstrated that VMDQ technology in Intel Network Adapters boosted max throughput on a 10gb virtualized connection from 4gb to ~9gb - nearly max theoretical capacity.

 

 

You can buy the bits today! There are about 30 different vendors with products in the 10gb space. Clearly this is a ripe area for innovation and entry. The question remains glass or copper - optical or electrical? There are pros and cons for each, and both are supported by this next generation Ethernet. I would love to hear which you are choosing and why.

 

 

 

 

0 Comments 0 References Permalink
0

 

Many utility companies in North America are encouraging energy efficiency in datacenters in a big way. Some are offering incentives to non-residential customers for making energy efficient choices including servers, storage, and other datacenter equipments. Most of these utility companies require customers to make thier rebate/incentive application prior to starting the project and obtain thier approval. I compiled a short informative article regarding various rebates/incentives offered by the utility companies and it was recently published in the datacenterjournal. Read through......

 

 

http://datacenterjournal.com/index.php?option=com_content&task=view&id=1475&Itemid=41

 

 

RK Hiremane

 

 

0 Comments 0 References Permalink
0

Today's workstation with (2) Intel® Xeon® 5400 Series processors transforms a workstation from a simple design terminal into a powerful engineering tool that helps users potentially compress the time between an idea and a product. At Intel we call that "working differently". In the manufacturing vertical market it may also be called:

  • digital prototyping,

  • analysis driven design or

  • simulation based product design.

 

No matter what you call it, these new ways of working help manufacturers virtually explore complete products before they are built-so they can create, validate, optimize, and manage designs from the conceptual design phase through the manufacturing process.

 

 

Digital prototyping processes may also require that you also rethink your workstation deployment strategy; moving from single processor workstations, to two (2) Intel Xeon 5400 processor based workstations with 8 computational cores and up to 100 peak gigaflops of floating point performance. This compute capacity coupled with the robust visualization environment enabled by two (2) PCI Gen 2 graphics adapters changes would be workstations into workstation supercomputers and enables engineers and designers to concurrently perform traditional CAD design as well moderate size analysis (e.g. over 5M degrees of freedom finite element modeling jobs or up to 10M cell fluid dynamic simulations).

 

 

This combination (CAD and Analysis) can help organization optimize CAD parts or assemblies under a wide range of physical scenarios including mechanical and thermal effects. Net result is workstation supercomputers based on Intel® Xeon® 5400 Series processor can help to bring higher quality or more innovative ideas to markets faster than ever before.

 

 

0 Comments 0 References Permalink

Filter Blog

By author: By date: By tag: