KVM-based Enterprise Virtualization 3.1 enables extra-large virtual machines and better live migration across more storage systems than before.
7 Cheap Cloud Storage Options
(click image for larger view and for slideshow)
Red Hat last week enhanced its open source alternative to Microsoft and VMware, Enterprise Virtualization 3.1, with the ability to mount larger virtual machines and achieve live migration across more storage systems than before.
It also cited the continued high performance of its kernel virtual machine, or open source KVM, in its Dec. 5 announcement.
Enterprise Virtualization 3.1 allows the creation of a virtual machine with up to 160 virtual CPUs and 2 TB of memory, said Red Hat's Chuck Dubuque, senior manager of product marketing. That's larger than the maximum 64 virtual CPUs and 1 TB of memory supported by Microsoft System Center's Virtual Machine Manager or VMware's vSphere 5.1.
The extra-large virtual machine that can run under Red Hat's 3.1 system means that "Enterprise Virtualization 3.1 can take the largest x86 boxes and virtualize them, those with eight CPU sockets with 10 cores each," said Dubuque.
Furthermore, these extra-large virtual machines are running the efficient KVM hypervisor, which uses the memory manager and scheduler inside the Linux kernel. That's part of the reason why KVM holds 19 of 27 published SPECvirt_sc2010 benchmarks in hypervisor performance. The SPECvirt is a general benchmark derived from an earlier VMware VMmark benchmark.
Results from a second independent benchmark, TPC-VMS, should be available soon to see whether its performance yields similar results for KVM. That benchmark comes from the Transaction Processing Performance Council.
Another new key capability is the 3.1 system's ability to live migrate virtual machines from one storage area network to another, using storage live migration. The feature matches what VMware and Microsoft can already do with their live migration systems, and it wasn't available in January's 3.0 release of Enterprise Virtualization. Red Hat's storage live migration "is in technical preview with 3.1 and will work," said Dubuque, but it is not a supported feature. It will be supported as it becomes generally available in a future release of the virtualization system, he said.
Red Hat added the storage live migration as a result of its October 2011 acquisition of Gluster, which created the open source GlusterFS storage file system. It was added to Red Hat Storage Server 2.0 in June this year, and allows Red Hat customers to scale out their storage systems across multiple storage domains. Enterprise Virtualization's ability to use Storage Server 2.0 capabilities reflects the first step in integrating the two, Dubuque said. That is, Red Hat plans to let customers take one x86 server and allow it work as both a virtualization compute and storage system.
Storage live migration is used when a large amount of data held on virtual disks must follow a migrating virtual machine, called vMotion in the VMware world. By including storage live migration, the virtual machine becomes independent of its underlying hardware and can migrate from rack to rack, across the data centers or even between data centers.
Dubuque said 3.1 allows live storage migration of a virtual machine's virtual disks from rack to rack and across the data center, but its current level of technical management doesn't allow transfers from one data center to another. In this release, storage live migration is "within a data center or where there is a significantly high bandwidth/low latency network between data centers," such as a campus-area network, he said.
Despite its limits, the beginning of storage live migration in Red Hat Enterprise Virtualization shows how the open source alternative is attempting to keep pace with its commercial counterparts.
Dubuque also noted that Enterprise Virtualization could previously access virtual images and data stored over iSCSI, Fibre Channel NFS or local storage. In the 3.1 version, it can access any storage managed by Storage Server 2.0, which means a wider set of options.
InformationWeek Elite 100Our data shows these innovators using digital technology in two key areas: providing better products and cutting costs. Almost half of them expect to introduce a new IT-led product this year, and 46% are using technology to make business processes more efficient.
The UC Infrastructure TrapWorries about subpar networks tanking unified communications programs could be valid: Thirty-one percent of respondents have rolled capabilities out to less than 10% of users vs. 21% delivering UC to 76% or more. Is low uptake a result of strained infrastructures delivering poor performance?
InformationWeek Tech Digest, Nov. 10, 2014Just 30% of respondents to our new survey say their companies are very or extremely effective at identifying critical data and analyzing it to make decisions, down from 42% in 2013. What gives?