Join me in Las Vegas in February for HIMSS 2012, the definitive Health Industry IT Conference. I will be working daily providing demos on how healthcare providers are leveraging self service portals, capacity on demand, and dynamic data centers to achieve a private cloud on VMware vSphere 5 and vCloud Director.
Join me at HIMSS Conference 2012
I am adding new topic to the Blog. High Performance Computing and Supercomputing on VMware vSphere. This topic came up from a recent customer discussion on how they could build a distributed cloud to the purpose of grid computing that scales on par with traditional super computers. Coming from a family that has been involved with supercomputers for decades, I found this topic to be of keen interest. As a result, I’ve started to collect a series of notes and white papers on the topic of High Performance Computing (HPC) on VMware vSphere and the changes that must be considered to maximize performance and scalability of a virtual HPC Cluster Environment.
First, an old but mandatory read on the topic is covered in a Paper published by Cam Macdonald and Paul Lau from the University of Alberta titled: Pragmatics of Virtual Machines for High-Performance Computing: A Quantitative Study of Basic Overheads by Cam Macdonell and Paul Lu
http://www.vmware.com/files/pdf/paullu.vmware.final.pdf
Macdonald and Lau’s paper is currently a little dated but many of the ideas still hold. Some points to be aware of is understanding how vSphere 5 has been optimized to reduce overhead, allow for better performance, and the automated orchestration features that will allow for capacity on demand. Macdonald’s and Lau’s closing issues regarding overhead have been reduced to less than 1% for most compute types and with the increase in performance of the CPU every year, this overhead becomes smaller yet.
Next, VMware Employee Jeff Buell posted a blog posts on “HPC Application Performance on ESX 4.1: Stream” back in Sept 2010.
http://blogs.vmware.com/performance/2010/09/hpc-application-performance-on-esx-41-stream.html
Jeff takes great strides in identifying factors that will optimize performance of the compute cluster. Most notable is the use of local memory when writing applications will ensure optimal memory bandwidth once deployed and keeping the computer resources within a single NUMA node to optimize resource utilization. While vSphere can address 1TB of RAM and up to 32CPU’s from a single VM, the optimization for performance lays on keeping VM’s tuned and sized to run within the optimal limitations of the server the VM is hosted upon.
Next, I want to make sure you follow the blog posts of Josh Simons. Josh works in the Office of the CTO at VMware as a strategist specializing in HPC and maintains the VMware Blog posts on HPC here: http://communities.vmware.com/community/vmtn/cto/high-performance
Josh has contributed several videos and discussions on the topics. With the recent Supercomputing 2011 event in Seattle, John pulled together several interviews and overviews of technologies that will enable Cloud based HPC.
In addition, Josh’s 2010 overview of HPC in the Cloud
Lastly, my own observations and comments:
With the recent release of vSphere 5 and Auto Deploy, the process of maintaining a scalable Cloud infrastructure has become considerably simplified. The process of updating an entire server farm can be reduced from weeks to minutes by leveraging PXE and an Image Server to refresh entire farms of servers at reboot. By adding solutions such as templates, workflow orchestration, and capacity management, we can now scale up clusters of computers on demand to accommodate almost any size distributed workload. Adding vCloud Director and vCloud connector allows us to scale the compute cluster even further into a single or multiple public cloud providers on demand. In addition, with the new scalability improvements of vSphere 5, we are finding larger VM’s, more addressable RAM, less overhead, and significant IO gains at the Hypervisor. All of these improvements contribute to the greater acceptance of HPC workloads in the Cloud and in a VM.
As I dig deeper into the topic, I hope I can contribute some of my own personal works to the field and leverage the knowledge of my colleagues to ensure others can explore this emerging growth area.
The wait is over. vSphere 5.0 was made available for download this evening by VMware.
Here are the highlights and links!
The VMware Licensing Advisor tool is now available for download. It will all you to connect to a vSphere 4.1, vSphere 4.0, or Virtual Infrastructure 3.5 environment to calculate the vRAM usage and vRAM capacity if the customer were to upgrade to vSphere 5.0.
The tool does require Java JRE 1.6 or Highrer
VMware has released its own License Validation Scriptfor vSphere 5.0 to address vRAM Entitlement. The Script provides a view of the revised (v2) vRAM Entitlements per vCenter. This script can be run against your current vSphere 4.x environment to see where you will land in the new model.
Note:
The script is available from VMware
Lightning Knocks Amazon, Microsoft Euro Clouds Offline | News & Opinion | PCMag.com.
An interesting article on how both Microsoft and Amazon EC2 were knocked offline by a lightning strike. All of the redundancies and generators could not help.
“While many volumes will be restored over the next several hours, we anticipate that it will take 24-48 hours until the process is completed,” Amazon added. “In some cases EC2 instances or EBS servers lost power before writes to their volumes were completely consistent. Because of this, in some cases we will provide customers with a recovery snapshot instead of restoring their volume so they can validate the health of their volumes before returning them to service.”
Amazon stated the following:
“Normally, upon dropping the utility power provided by the transformer, electrical load would be seamlessly picked up by backup generators.”
“The transient electric deviation caused by the explosion was large enough that it propagated to a portion of the phase control system that synchronizes the backup generator plant, disabling some of them.”
From Microsoft:
“We were informed the incident was the result of an lightning strike that caused an explosion on one of the substation’s transformers. The lightning strike created an electrical surge large enough that it affected a portion of our backup power systems. Our team is highly trained to respond to various types of issues, so they then began to manually transfer to generator power.”
Looks like VMware made public a revised vSphere 5.0 License Model today. A lot of very good changes that directly benefit the end user. ( http://www.vmware.com/files/pdf/vsphere_pricing.pdf )
The biggest changes are around vRAM Maximums per processor.
vRAM Entitlement per CPU Socket by vSphere edition
- 32GB vRAM/CPU for Essentials Kit (up from 24GB)
- 32GB vRAM/CPU for Essentials Plus Kit (up from 24GB)
- 32GB vRAM/CPU for Standard (up from 24GB)
- 64GB vRAM/CPU for Enterprise (up from 32GB)
- 96GB vRAM/CPU for Enterprise Plus (up from 48GB)
This means a 4CPU Server of Enterprise Plus is entitled to 384GB of powered on VM’s, a 100% increase over the initial model. These entitlements can also still be pooled across vCenter Server (including vCenter Servers in Linked Mode) and only apply to the actual configured virtual RAM of powered on VM’s. VDI Users and VMware View users have a different license model for High Density VDI servers. (Discussed below)
VMware Server (the free hypervisor) users will also be happier. Under the old model, VMware Server was capped at 8GB. The revisions have increased the cap to 32GB per server (not a per CPU limit). This is a hard limit and cannot be exceed like the retail versions. Personally, I felt the 8GB limit was very limiting. The 32GB change is a significant step in the right direction and should accommodation most SMB’s/Branch Office/Home Office users. Those that feel hindered by the 32GB hard limit should look at vSphere Essentials and Essentials Plus (starting at $560) which provides 6 CPU’s (up to three servers with 2CPU’s each) with a vRAM entitlement of 192GB. Alternative Options are to move to the revised vSphere Acceleration Kits (Standard, Enterprise, and Enterprise Plus) which provide bundles of 6 CPU’s at substantial discounts. Pricing has vSphere Essential starts at $83/CPU retail!
Even better, There is no longer a penalty for very large VM’s. All VM’s only count vRAM consumption up to 96GB. Any vRAM over 96GB is not counted. That means a 1TB VM would be covered by a single CPU license. A 1TB VM is now covered by a single Enterprise Plus CPU server license. I can’t imagine running more than a single 1TB VM on a host, but people will always surprise you.
vRAM usage is now monitored on a 12 month rolling average with daily high water marks. This makes large infrequent deployments less of an issue for customer who anticipate going over the vRAM entitlement but know that they will be removing VM’s later.
As before, VMware View environments don’t follow the vRAM model. View is a CPU Socket based license for View Desktops. Non-VMware View users will be able to leverage a vSphere for Desktops product just for VDI. vSphere Desktop edition is licensed based on the total number of Powered On Desktop Virtual Machines and can be purchased either stand-alone in a pack size of 100 desktop VM or included with the VMware View Bundle.
Another interesting note for current vSphere 4.1 users with valid Support and Subscription (SnS). Customers who purchased licenses for vSphere 4.x (or previous versions) prior to September 30, 2011 to host desktop virtualization, and hold current SnS agreements, may upgrade to vSphere 5.0 while retaining access to unlimited vRAM entitlement. Desktop licenses covered by this provision, however, may not be managed by the same instance of Virtual Center which is being used to manage non-desktop OS virtual machines.
Lastly, I have heard that vCenter will get an update in the near future after release to accurately report these last minute changes. In the mean time we should be expecting a new tool to report the actual vRAM consumption.
There are lots of key aspects that were addressed in the initial vSphere 5 guide that have not changed but are different form the vSphere 4 model. Over all, I think the consumer gains a lot more value in the new editions! Unlimited RAM capabilities per server, no more Core/CPU limits, substantial increase in CPU’s per VM limit by edition (32 CPUs/VM for Ent+!) and vMotion all the way down to the Standard Edition.
A lot has changed. I see it all as a major improvement for both the SMB users and for the Enterprise.
Taken from revised the vSphere 5 Pricing Guide
I was not a fan of the Apple Linen Login screen and decided to dig around to see what was needed to replace it with something more suited for my tastes. This is my new login screen and logo.
Apple OS X Lion Login Screen Replacement. New background and Logo. Wallpaper by Thomas Sychra http://www.zpecter.net/
How to Replace the Apple Logo
First, you will need a new transparent .PNG File. I created a VMware Logo from an existing corporate file, captured, removed the background, and saved as a transparent .PNG file.
There are two logos and they differ a little depending on the background. I created a White VMware Logo and saved it for both images.
Rename the old images to .old and then drop in the new images with the same old names of apple.png and appleLinen.png. Be sure they are in a .PNG file format. A benefit is that .PNG allows for transparency.
Path:
/System/Library/PrivateFrameworks/LoginUIKit.framework/Versions/A/Framework s/LoginUICore.framework/Versions/A/Resources
File:
appleLinen.png
apple.png
How to replace the login page background
These are System files and will require administrative rights to replace or rename.
They are stored at the following directory path
Path:
/System/Library/Frameworks/AppKit.framework/Versions/C/Resources/
File:
NSTexturedFullScreenBackgroundColor.png
The background is also a .PNG image (1280×800 @ 72dpi for my 13″ Macbook)
Rename the old image to .old to keep the original and then drop in your replacement image. Images smaller than your screen resolution will automatically tile.
Ideas:
- Add a wall paper that has your contact information in the event you ever lose your laptop or if it is stolen.
If you use this technique or you find this helpful, drop me a note at chris@grossmeier.com or leave a comment below!
Hugo Peeters of the Netherlands blog PeetersOnline.nl has written an excellent script to run against your vCenter to calculate your new vSphere 5 vRAM License details. So far, I have not seen a single environment across my customers who don’t already have far more vRAM entitled to them than they are using. This little PowerCLI script will poll your servers, collect the pCPU Count, look at the editions, add up your current vRAM consumption and comapre it to your entitled vRAM. An elegant script that dispells the idea that customers are using 100% of their pRAM for vRAM across the datacenter. I have even seen 256GB+ VM’s easily fall within the vRAM of the vCenter’s pool sizes. Remember,
Get the script and run it! Great job Hugo!
URL : http://www.peetersonline.nl/index.php/vmware/calculate-vsphere-5-licenses-with-powershell/
UPDATE: Virtu-AL also has a different script that provide similar information.
http://www.virtu-al.net/2011/07/14/vsphere-5-license-entitlements/
I want to share something on licensing I have been discussing others who hear about vRAM but never took the time to read what the new product licensing entitlement. (You can read the details in the VMware vSphere 5.0 Licensing, Pricing and Packaging White Paper)
I saw this correlation earlier today from someone else….
“Think of it this way. We just invented electricity. Everyone loves it, it makes them more efficient, safer, etc. As the power company, we initially decide to charge for our electricity based on the size of the house. Seems fair at first, but what if one two bedroom home powers 20 lights and another 2 bedroom home powers 200 lights, a pool, 3 freezers, etc. Same physically sized environment, totally different consumption level. Is it fair for the 1st house to pay as much as the 2nd for our electricity? Maybe the most fair way would be to charge per light (per VM), but would customers want to keep track of that and be nickel-and-dimed for every new light they turn on? “ – Jason Dion
I expanded the analogy as follows:
The future is to enable the enterprise to turn virtualization in to a utility service which is better suited for services such as Cloud Computing, capacity on demand, and dynamic growth. Processors are not a stable metric to base this type of model on since the processor your VM runs on may be outside of your own managed datacenter (the Hybrid Model). With vSphere 5, vRAM, memory consumed by a virtual machine, is a less nebulous entity that remains the same regardless of growth or location of compute. What happens when Intel or AMD release a new large-number core Processor greater than 12. What if you’re not using all the pRAM in that 256GB Server and only had 8GB of pRAM consumed? vRAM is a known and predefinable entity that leverages a more fair use model. In addition, you gain the benefit of aggregated vRAM across resources! You are not limited by the same physical boundaries of a single server node.
Here is what I shared with individuals from the local VMUG who had questions on the topic…
vRAM is not so much a restriction as it is a liberator form the traditional Datacenter licensing. No more Cores/VM or max system RAM limits. Also, one clarification… vRAM is consumed RAM across the pool of ESX hosts managed by vCenter, not how much RAM in in the Server. vRAM is aggregated across the vCenter and is the amount of consumed RAM across all ESX hosts in a vCenter (Good news, Linked Mode counts too). You get to pool your RAM across servers to define your available vRAM. It is not a per server number.
The following statement is direct from VMware’s Marketing Team:
VMware is evolving the product’s licensing to lay the foundation for customers to adopt a more “cloud-like” IT cost model based on consumption and value rather than physical components and capacity. VMware vSphere 5 will continue to be licensed per processor (CPU), however, VMware is eliminating the current, restrictive physical entitlements of CPU cores and physical RAM per server and replacing them with a single, virtualization-based entitlement of pooled virtual memory, or vRAM.
Pooled vRAM is the total amount of memory configured to all VMs in a customer’s environment. Each VMware vSphere 5 CPU license will entitle the purchaser to a specific amount of vRAM, which can be pooled across the entire vSphere environment to enable a true cloud or utility based IT consumption model. There are no restrictions on how vRAM capacity can be distributed among VMs: a customer can configure many small VMs or one large VM. VMware vSphere has made it possible for customers to maximize hardware utilization and efficiency by pooling CPU, memory, storage and networking. With these licensing changes, VMware is extending the concept of pooling – one of the foundational elements of cloud computing – beyond technology to the business, allowing the pooling of licenses for maximum utilization and value.
Update:
See my post on Scripts to show you your Consumed vRAM and vSphere vRAM entitlement.
