Cost versus Performance... The battle continues...

Over the past couple of weeks, I've had various meetings discussing upcoming features and roadmaps for both the latest versions of VMware 4.x and vCenter, as well as Hyper-V 2008 R2 and SCVMM. 

The battle continues between Microsoft's debate on cost savings versus VMware's claim to fame that their product is not only more cost effective (per app), but their expertise in virtualization performance coupled with their advanced capabilities that Hyper-V has yet to match, keeps them on top as an enterprise virtualization solution. The question is... How close is Microsoft to either A. Matching the capability that VMware has (if ever)? and B. Which solution is more important to your organization, cost or performance?

Microsoft in June 2010 conducted a very extensive study (though I have to assume that the metrics are slightly skewed) on the cost benefit analysis of Hyper-V to VMware.

The key points in the cost analysis was that Microsoft has a cheaper solution when you break it down per guest operating system as well as a cheaper solution when looking at annual labor costs within a VI... Though the one thing that it doesn't seem to discuss, is not the performance of the VI test model, nor did it discuss existing capabilities within the products that were tested. Knowing that this was a cost analysis and only a cost analysis I think that Microsoft could in fact be a cheaper (bottom line dollar amount) product... Meaning in test lab environments it's an ideal solution (as you can see below)...

Workload Comparison – Average Cost per Guest VM by Platform

The respondents were asked to identify the workloads their company had virtualized. Across workloads (App, DB, DR, Email, Test & VDI) the Hyper-V costs were less expensive for all workloads except test – in all other workloads the Hyper-V customer saw significant savings.

                                                                                                  (Microsoft, 2010).

The bottom line is looking at the above, that most platforms are utilizing VMware versus Hyper-V (with the exception of test/lab environments). If you look at some of the noteworthy platforms (VDI, E-mail, and App), you can see that VMware is even more heavily utilized (though costing more). I believe this to be true because these environments are extensive to configure in a VI (especially a VDI solution). Also, because VMware has more redundancy measures you can configure within those environments, the labor associated with the design and operations may also be increased.

Bottom line, it's the organization that decides the benefit of their VI design and layout (be it cost, features, performance, or a combination of all of the above based on the requirements). However, I think David Davis a contributor to SearchVMware.com nailed it on the head...

Why VMware ESX beats Microsoft Hyper-V, hands-down

In the end, it is my opinion that VMware "wins the war" for several reasons. Perhaps most obvious, is that Microsoft is already incredibly behind VMware in terms of virtualization know-how and may never catch up. In 2007 alone, VMware announced ESXi, Site Recovery and Update Manager in an effort improve ESX Server. As they will continue to improve their product year after year to provide more value, Microsoft is fighting an uphill battle.

In addition, VMware will continue to see a huge surge in revenue thanks to Virtual Desktop Infrastructure (VDI). VDI, in my opinion, will be the next wave of virtualization after server consolidation. Also, VMware will be able to compete in the price war with Microsoft because of ESXi. The company has the option of giving it away for free, but even if they don't they're still offering it at $495 as opposed to Windows Server 2008 Hyper-V at $999.

 

Finally, and most importantly, ESXi can still win in efficiency. Compare the ESXi hypervisor at 32MB to the size of Hyper-V at about 2GB. Who will win at boot up? Who will win at security? Who will win when it comes to a lack of troubleshooting and solid reliability? My guess is that 32MB worth of code is going to be inherently faster, more secure and more reliable. In terms of reliability and uptime, it has been stated that ESX will run for 1000+ days without a reboot, whereas Windows Server 2008 Hyper-V will need to be rebooted every 30 days due to Windows Updates.

 

In the end, if you look at just a single license of VMware Infrastructure Suite Enterprise at $6950, and compare that to a single license of Windows Server 2008 Enterprise or Standard, the cost for the VMware solution will undoubtedly cost more. However, it still beats Microsoft's Hyper-V in terms of performance hands-down.

                                                                                             (Davis, D., 2008).

Regardless, of your environment and budget, requirements are going to define whether or not you utilize Hyper-V or VMware. If you have no requirement to deploy a highly available solution and you have a limited budget, then Hyper-V is the answer. If you have a large environment with a complex storage and network implementation and you want to utilize redundant solutions within your host, guest OS's, and network implementation then VMware with vCenter is the answer.

On a personal note, I hope everyone has a Happy Holiday and a Happy New Year... More blogs to come in 2011. Hope you enjoy!

References:

Davis, D. (2008). How VMware ESX performance trumps Hyper-V's price. Retrieved on December 21, 2010 from How VMware ESX performance trumps Hyper-V's price

Microsoft. (2010). Microsoft Hyper-V vs. VMware ESX & vSphere Operations & Management Cost Analysis. Retrieved on December 21, 2010 from Microsoft Hyper-V vs. VMware ESX & vSphere Operations & Management Cost Analysis

Finally nice to see some serious competition in the virtualization marketplace...

Watch out VMware, because Microsoft is going to catch up and provide some serious competition in the virtualization and cloud computing marketplace!!!

I've had the recent privilege of playing with System Center Virtual Machine Manager (SCVMM) from Microsoft (which is part of the System Center Suite) and got to manage some physical Hyper-V hosts. I have to say that with R2, the functionality of SCVMM is comparable to vCenter. They are working on developing the methods to match VMware's capability of amazing functions like Storage vMotion, vMotion, DRS, and HA.

What's even crazier is that SCVMM can manage vCenter instances (limited capabilities as opposed to using vCenter) and ESX/ESXi hosts.

Now these features (almost all of them) are new to R2. Meaning six months ago, VMware was the only decent virtualization product out there that supported serious clustering models for a larger enterprise. Witih the release of SCVMM 2008 R2, there is definitely some competition brewing between Microsoft and VMware.

One area that is difficult to negotate (for many businesses) is the cost of licensing. Microsoft has VMware beat (to a certain degree) with their Datacenter licensing model. Server Management Suite Datacenter (SMSD) costs about 1,500 per 2 procs and you have to purchase it at a 2 proc minimum; whereas vCenter alone is about triple the costs and Enterprise Plus licenses is about 3,900 MSRP per proc.

However, as it stands right now, VMware still offers more networking and advanced storage options. Though this leaves me to believe that with R2's included migration, networking, and current storage options now, Microsoft isn't too far behind in being a viable competing organization in the virtualization and cloud computing market. Especially, given SCVMM's ability to integrate with the entire System Center productivity suite (Operations Manager and Configuration Manager). I hope to receive more formal training on SCVMM and VMware in the next few months so I'll be able to provide a more educated comparison of the two products in later blogs. For now, check out the new features of SCVMM 2008 R2.

What's New in VMM 2008 R2

System Center Virtual Machine Manager 2008 (VMM 2008) is a comprehensive management solution for managing virtualized infrastructure running on Windows Server 2008 with Hyper-V, Virtual Server 2005 R2 and VMware ESX through Virtual Center. Recently, Windows Server 2008 R2 RC was released, which includes significant feature improvements to Hyper-V—the underlying hypervisor platform. VMM 2008 R2 – the next version of VMM – has recently been released. VMM 2008 R2 leverages the new platform enhancements and extends the feature set of VMM 2008. This overview highlights the most important new and significantly enhanced features in the VMM 2008 R2:

Support for new features of Windows Server 2008 R2

• Live Migration: Seen through the VMM console, this enables administrators to move a virtual machine between clustered hosts in a way that is completely transparent to the users connected to the virtual machine. This allows administrators greater flexibility in responding to planned downtime and provides higher machine availability. The basic requirements for Live Migration are that all hosts must be part of a Windows Server 2008 R2 failover cluster and host processors must be from the same manufacturer. Additionally all hosts in the cluster must have access to shared storage. No changes are required to existing virtual machines, network, or storage devices in moving from Quick Migration to Live Migration other than upgrading to Windows Server 2008 R2 and VMM 2008 R2.

• Hot addition/removal of Storage: Allows the addition and removal of storage to virtualized infrastructure without interruption. Additionally, "live” management of virtual hard disk (VHDs) or iSCSI pass through disks, allows administrators to take advantage of additional backup scenarios and readily use mission critical and storage-intensive applications.

• New optimized networking technologies: VMM 2008 R2 supports two new networking technologies – Virtual Machine Queue (VMQ) and TCP Chimney – providing increased network performance while creating less of a CPU burden. NICs that support VMQ, create a unique virtual network queue for each virtual machine on a host that can pass network packets directly from the hypervisor to the virtual machine. This increases throughput as it bypasses much of the processing normally required by the virtualization stack. With TCP Chimney, TCP/IP traffic can be offloaded to a physical NIC on the host computer reducing CPU load and improving network performance.

Enhanced storage and cluster support

• Clustered Shared Volumes (CSV): Provides a single, consistent storage space that allows hosts in a cluster to concurrently access virtual machine files on a single shared logical unit number (LUN). CSV eliminates the previous one virtual machine per LUN restriction and coordinates the use of storage with much greater efficiency and higher performance. CSV enables the Live Migration of virtual machines without impacting other virtual machines sharing the same LUN. Enabling CSV on failover clusters is straightforward; many storage configuration complexities prior to CSV have now been eliminated.

• SAN migration into and out of clustered hosts: This allows virtual machines to migrate into and out of clusters using a SAN transfer, which saves the time required for copying the virtual machine file over the network.

• Expanded Support for iSCSI SANs: Previously, only one LUN could be bound to a single iSCSI target whereas now – with support now built into VMM 2008 R2 – multiple LUNS can be mapped to a single iSCSI target. This provides broader industry support for iSCSI SANs allowing customers more flexibility in choosing storage providers and iSCSI SAN options.

• Storage Migration: Quick Storage Migration enables migration of a VM’s storage both within the same host and across hosts while the VM is running with a minimum of downtime, typically less than 2 minutes. VMM 2008 R2 also supports VMware storage vMotion which allows the storage of a VMware VM to be transferred while the VM remains on the same host with no downtime.

• Rapid Provisioning: Allows administrators to take advantage of SAN provider technologies to clone a LUN containing a VHD and present it to the host while still utilizing the VMM template so the OS customization and IC installation can be applied.

• Support for third party CFS: For users requiring a true clustered file system, VMM 2008 R2 supports third party file systems by detecting CFS disks and allows for deploying multiple VMs per LUN.

• Support for Veritas Volume Manager: VMM 2008 R2 recognizes Veritas Volume Manager disks as a cluster disk resource.

Streamlined process for managing host upgrades

• Maintenance Mode: Allows administrators to apply updates or perform maintenance on a host server by safely evacuating all virtual machines to other hosts on a cluster. Maintenance mode can be configured to use Live Migration to move the virtual machines or can put the workloads into a saved state to be safely reactivated when maintenance or upgrades are complete. Maintenance mode is enabled for all supported hypervisor platforms on Windows Server 2008 R2.

Other VMM 2008 R2 enhancements

• Support of disjoint domains: Reduces the complexity of reconciling host servers with differing domain names in Active Directory and DNS. In these situations, VMM 2008 R2 automatically creates a custom service principal name (SPN) configured in both AD and DNS allowing for successful authentication.

• Use of defined port groups with VMware Virtual Center: On installation, VMM 2008 R2 will present available port groups for VMM’s use with VMware vCenter thus allowing administrators to maintain control over which port groups are used.

• Queuing of Live migrations: This feature enables users to do multiple Live Migrations without needing to keep track of other Live Migrations that are happening within the cluster. Detects when a Live Migration will fail due to another Live Migration already in progress and queues the request for later.

• Host compatibility checks: VM migration requires host hardware to be compatible; this feature provides a deep check for compatibility using Hyper-V and VMware compatibility check APIs. Administrators can check if the source host is compatible with the destination host before performing a migration and finding out the VM cannot start on the new host. A related feature makes a VM compatible by turning off certain CPU features which makes the VM compatible with the hosts in the cluster.

                                                                                      (Microsoft, 2010).


Reference

Microsoft (2010). What's New in VMM 2008 R2. Retrieved on December 1, 2010 from
           What's New in VMM 2008 R2 - Microsoft Corp.

A very in depth comparison when choosing virtualization platforms...

I've been getting into discussions with my Microsoft Premier Field Engineers at work about VMware and Hyper-V. I won't argue that Hyper-V is cheaper (requires just the OS license from Microsoft and then you can load Hyper-V onto it). However, I get caught up in features and the old saying that Microsoft never gets anything right until the third time...Regardless, I always try to give any vendor the benefit of the doubt and always approach a new delivery from not just a technical perspective (comparison of features, expertise, etc...) but also a business one (costs versus time). I won't disagree that VMware isn't expensive Enterprise Plus is about 3,800 per processor not including costs for vCenter. However, I stumbled across the ultimate comparison guide between the four major vendors (VMware, Microsoft, Citrix, and Redhat). If this doesn't provide you with enough information about which virtualization vendor to choose from, I'm not sure what will... Enjoy

Now if there was ever a reason to purchase VMware over the other, the below information shows features and options offered by VMware that aren't offered (at least fully featured) by the other vendors. Of course this comparison was conducted by VMware so take it with a grain of salt. However, no comparison offered by a particular vendor would be complete without a selling point about why you should go with their product.

As with anything you are building into your infrastructure, you need to outline business and technical requirements and feasibility before engineering a vendor based solution, especially when it comes to virtualization.

Here is a link to the full PDF file with the above information...

VMware vSphere—The Best Platform for Building Cloud Infrastructures

Is there such thing as too high available?

Some of the amazing features of vSphere v4.x is the added availability and redundancy of resources managed by vCenter.

vCenter is the management product used to manage VMware VMs, vNetworking, Storage, and more....

There are two things that you need to have in place in order for many of the available operations of VMware to function. The most important one is the licensing to support it. Most large organizations are paying for Enterprise or Enterprise Plus for their Virtual Infrastructure (VI), for those that pay for standard or advanced you are limited to the features you can use. A breakdown of the licensing model is below...

                                          

 (VMware, 2010).                                                                                   

The other thing is the resources in your cluster to support the neat features that expand the availability and redundancy of your VI. This is very important as the larger your cluster (in terms of hosts) the more redundant your VI is without added features with vCenter, you can simply use High Availability (HA) coupled with Distributed Resource Scheduling (DRS) to ensure that your environment is not only utilizing resources in a balanced and efficient manner, but also that VMs can be moved to other hosts.

This is common for those running enterprise licenses or greater. For those running advanced or fewer, you just have the high availability function (advanced users have vMotion as well which performs live migrations of your VMs to other hosts with zero downtime)

This is just one option and one scenario. Other products like Fault Tolerance (VMware's Clustering product with other VMs in a cluster), vCenter Server Heartbeat (create a vCenter active/passive cluster), and Distributed Power Manager (this really isn't a high available option but is a great product to efficiently use your hosts).

What's really neat about all of the above mentioned products is that they can be used in combination with one another (given the appropriate licensing levels). For example, use DRS with vMotion and you'll not only have live migration but efficient balancing of resources in your cluster.

Now if you want to migrate files between datastores so you can patch a given storage device or appliance, you have Storage vMotion to allow you to do live migration of VM's (files) from one store to the next.

So the point of this is how redundant or high available do you want your systems to be? To me, it would be where you draw the line between effective and efficient utilization versus time spent maintaining the systems. If your environment has critical VMs that require 100% uptime due to critical SLA's then I'd say that you have a requirement to maintain the VI at the highest available level (vMotion with DRS in a FT cluster where vCenter is using Heartbeat). If not, then doing something simpler while providing that level of availability (perhaps HA only). This allows recoverability of VMs though you may experience minimal downtimes. Below is a layout on the migration and high availability products commonly used with VMware and how it is broken down.

                                                                                                                                         (VMware, 2010)

                                                                                                                                          (VMware,2010).

These products are all meant to be used to make VMware environment with vSphere more available. Additionally, they are also broken down to be geared towards anyone that could use any given module based on it's cost. i.e. Small Business owners could easily afford vSphere Standard or Advanced to use vMotion or High Availability. Where your corporate enterprises could easily affort Enterprise or Enterprise Plus and use the added functionality of DRS and even purchase Heartbeat if needed. This product is robust enough to allow you to mold your VI based on not just your requirement (say an SLA) for uptime, but your engineering and administration team's ease of mind knowing that the VI is easily recoverable or highly available so there won't be that wake up call at 3AM as often...

Yes I am... and don't call me virtually...

Seriously though... Virtualization technologies have been around for many years now. Yet many small businesses have yet to still even consider virtualization into their environments. Is it the cost? The cut-over? The lack of knowledge? Support? All of these things probably have some merit into why organizations have yet to consider or even look at products like HyperV, VMware, and Sun Virtual Box to help consolidate their power, space, and cooling issues in their data centers and help create what could be centrally managed virtual devices that run on larger pieces of hardware resources. Maybe it's that concept that escapes many businesses and puts them in a state of fear that prevent them from even considering the concept of virtualization.

Recently, I had the privilege of taking the VMware Fast Track course (my second VMware course this year) on the vSphere 4.0 instance of VMware. There I learned about what was new in vCenter, Networks, Storage, and all the inner workings of ESX and ESXi. Here, I saw some really interesting points that discussed key differences between bare-metal versus virtual systems.

- Physical Machines are difficult to move or copy, are bound to specific set of hardware components, often have a short lifespan (about 3 years), and requires direct contact to upgrade the hardware (which impacts that systems availability).

- Virtual Machines are easy to move and copy between physical hosts because they are encapsulated into files (which are easy to transport) and independent of the hardware resources available. They are easy to manage because they get isolated from other virtual machines running on that same physical host and are insulated from physical hardware changes.

This makes me wonder, and I can't confirm this but is my opinion nonetheless, but when blade technologies from Dell, HP, Sun, Cisco, etc... were built, virtualization really had a means to take off because of the scalability of these systems and the fact that with these technologies we could consolidate all our hardware resources into a single pool. From here this is where I believe VMware really took off and ran away with the virtualization market with the development of vCenter.

vCenter is a centrally managed software device that runs on a Windows operating system that is used to managed multiple hosts and virtual machines. It is also used to manage other hardware resources available to your data center (physical hosts that fall under your instance of vCenter). I'll get into the finer tuned details of vCenter in future blogs.


(VMware, 2010) - Edited by DHood to include VM layers



*This blog is meant to provide a fundamental understanding of virtualization technologies (more specifically VMware). The views and opinions expressed in this blog do not reflect those of VMware or any other technology vendor.