LAS VEGAS -- After Palo Alto, Calif.-based VMware Inc. announced a new fault tolerance feature to be included in...
By submitting your email address, you agree to receive emails regarding relevant topic offers from TechTarget and its partners. You can withdraw your consent at any time. Contact TechTarget at 275 Grove Street, Newton, MA.
Virtual Datacenter Operating System (VDC-OS), a software company that lives and breathes fault tolerance with its own everRun technology gave a harsh critique.
In response to VMware's new fault tolerance feature, which was demonstrated during a VMworld keynote, Littleton, Mass.-based fault tolerance software developer Marathon Technologies published a blog knocking VMware's attempt at fault tolerance.
In 2004, Marathon began offering fault-tolerance software that works with standard off-the-shelf x86 Intel and AMD servers with Windows Server 2003 and unmodified Windows applications and then, in March 2008, it began offering fault tolerance software for Citrix XenServer virtual machines (VMs).Marathon critiques VMware fault tolerance
"While not bad for FT rookies, from what we saw, it's a less-than-perfect solution for a lot companies that want to run business-critical and mission-critical applications in VMs," Marathon stated.
And at the VMware-sponsored VMworld 2008 conference in Las Vegas, Marathon's senior director of product marketing, Michael Bilancieri, compared VMware's version of fault tolerance with its own.Marathon Technologies' Michael Bilancieri compares everRun with VMware's fault tolerance technology
According to Marathon's blog posted by Brian Mullins, the senior director, corporate communications, VMware's fault tolerance falls short in three key areas: component-level fault tolerance, complexity and limited CPU fault tolerance.
"The most common failures that result in unplanned downtime are component failures such as storage, NIC [network interface card] or controller failures. Yet VMware Fault Tolerance doesn't do anything to protect against I/O, storage or network failures," Mullins wrote. "By not addressing these primary sources of failures, VMware appears to be saying that you/the customer are on your own do figure out how to protect your storage and network connections. This may be okay for the very largest IT staffs in the world, but for the other 98%; it will not be sufficient."
As for complexity, Marathon complains that in order to use VMware fault tolerance, users have to install both VMware High Availability and Distributed Resource Scheduler (DRS). VMware fault tolerance also requires NIC teaming, so users have to manually install paired NICs and manually set up dual storage controllers (with the software to manage them), because it requires multipathing. "And to top it all off, you're required to use an expensive, and often complicated, SAN [storage area network]."
Also, according to Marathon, with VMware FT a "record/replay" capability has to be set up on both a primary and secondary server. "If something happens to the primary server, the record is stored on the SAN and then restarted on the secondary server. "First, the whole thing depends on the quality of the SAN. Second, in the words of the VMware engineer who presented at VMworld, "this can take a couple of seconds." So what happens to your application state in those couple of seconds?
As Marathon makes clear, another issue is VMware's fault tolerance applies only to VMware environments. That said, it should be noted that Marathon does not support VMware or Microsoft Hyper-V as of yet.
During his keynote address on Sept. 17, VMware's chief technology officer, Stephen Herrod, discussed this fault tolerance feature.
Herrod and Mark Vaghn of the First American Corp., who has used the test version of VMware's fault tolerance feature, demonstrated the new feature. In the spirit of Las Vegas, Vaghn ran a virtual slot machine called Cherry Bandit and Herrod "mistakenly" turned off the physical server hosting the VM running that application. As far as the audience could see, the slot machine wheels kept spinning when the server went down because the fault tolerance VM kicked on in lockstep.
"With fault tolerance, you are running a VM and a shadow VM, so if the physical hardware dies, the shadow VM kicks on and the application continues running without interruption," Herrod said. Following that, a shadow of the shadow VM is automatically created to protect it from downtime, he said.
Bilancieri admitted that VMware's ability to make shadow copies of shadow VMs when the primary VM goes down is impressive because it ensures the backup VM is protected.
"That is one area where VMware has done something interesting. Their product is integrated in with their DRS capabilities, which allows them to do [shadows of shadow VMs]," Bilancieri said. "It is something we are working on and will be included in future versions of everRun VM."
Marathon does not support VMware, and Bilancieri could not discuss why other than alluding to some "discussions" with the company. Marathon does plan to support Microsoft Hyper-V in the future though.
For more on the VMworld 2008 conference, check out our VMworld 2008 conference page.
Dig Deeper on Disaster recovery, failover and high availability for virtual servers