Red Hat is emphasizing KVM’s scale-up capabilities with today’s beta release of Red Hat Enterprise Virtualization 3.0.
The updated KVM hypervisor supported with Red Hat Enterprise Virtualization (RHEV) 3.0 is based on Red Hat Enterprise Linux (RHEL) 6.2. Among its capabilities is support for up to 2 TB of memory and up to 64 virtual CPUs per virtual machine (VM). By contrast, VMware’s forthcoming vSphere 5 will support up to 1 TB of memory and 32 CPUs per VM.
Red Hat is clearly trying to take advantage of the uproar over VMware’s vRAM licensing, which charges vSphere 5 users based on the amount of virtual memory allocated out of a pool to VMs in a cluster. By contrast, Red Hat charges a per-socket, per-year subscription fee for software support and does not limit the size of the memory pool.
(Red Hat’s blow is somewhat softened by the fact that VMware recently changed its vRAM licensing terms to make them more palatable for users and instituted a cap so that users never pay for more than 96 GB of vRAM per VM.)
Red Hat claims KVM is gaining momentum among users who are looking for an alternative and concerned about being locked in with a proprietary vendor, i.e., VMware. And the company has formed a marketing coalition called the Open Virtualization Alliance, which also targets VMware, to promote KVM’s usage. However, Red Hat is not releasing any specifics on the number of customers for RHEV or KVM to date, or growth in adoption of either technology.
Red Hat officials also declined to give specifics around RHEV 3.0 support for scale-out infrastructures while the product is still in beta. The maximum number of hosts supported per KVM cluster is still being defined, said Navin Thadani, senior director of Red Hat’s virtualization business.
Beta release dumps Windows, brushes up GUI
Some RHEV 3.0 features, announced at this year’s Red Hat Summit, are turning the heads of customers already in the KVM camp — especially the fact that the RHEV management server (RHEV-M) no longer has to run on a Windows host. Instead, RHEV-M now runs as a JBoss application on RHEL 6.
The RHEV-M user interface will now offer the ability to perform role-based administration and a graphical mapping of resources, such as storage devices, for easier management. It also includes historical performance reporting based on the open source JasperReports. Another addition to RHEV 3.0 is support for storage directly attached to a server; previously, RHEV required shared storage. And there’s a new generic end-user self-service portal, so customers don’t have to roll their own.
RHEV 3.0 will also support a new RESTful API against which users can write custom scripts to add advanced features. The features come from the command-line based KVM management utility libvirt, which can be called by the RHEV API and execute functions such as CPU pinning, single-root I/O virtualization, direct LUN access from the virtual host, putting switch ports into promiscuous mode for monitoring, watchdog monitoring and integration with Cisco’s VN-Link. These features are not yet part of the RHEV GUI.
These developments had users who’ve already bought in to RHEV excited at the Summit conference, and attendees said that they are evaluating Red Hat’s latest virtualization products. But production deployments were scarce, with users saying KVM is fighting an uphill battle against incumbent hypervisors. Similarly, the OVA is adding partners quickly, but Linux users also say it will take more than a marketing alliance to boost adoption of KVM over VMware’s ESX and ESXi.