While it’s true that this a 1.0 version and it’s then acceptable that it has a number of limitations, it’s still true that Red Hat is supposedly developing this product since at least one year and a half (as soon as it acquired the startup Qumranet) and that it has to immediately deliver a very competitive product if it wants to play against VMware.
And probably to accelerate the process of maturation of RHEV, this week the company announced that Red Hat Enterprise Virtualization Manager (RHEV-M) will become an open source project.
Unfortunately this won’t help still for a long time.
In a recent online presentation in fact, Red Hat revealed a number of details about its roadmap for RHEV that are all but promising:
- the upcoming RHEV-M 2.2 will still require Windows Server 2008
- to offer RHEV-M high-availability then, customers will have to use Windows clustering services
- while the product is being ported now in Java so to be used on every platform including Linux, the new Java-Based RHEV-M won’t be ready before another year or so (probably in the form of version 3.0)
- RHEV-M 2.2 probably won’t support any form of storage live migration
On top of that, apparently, Red Hat has no current plans to offer a limited free edition of its KVM-based hypervisor as VMware ESXi, Citrix XenServer or Microsoft Hyper-V are.
Despite that, and the many limitations above, Red Hat is claiming a significant saving when its virtualization platform is compared to VMware and Microsoft ones:
(quite interestingly, Citrix is not part of the comparison)
At the same time Red Hat claims better performance than competitors with several mission critical workloads:
Easy to guess, VMware will have something to say about this.
Update: Apparently, the RHEV-M porting to Java will imply the adoption of JBoss.
Support for third party databases (like Microsoft SQL Server and PostgreSQL) is also planned.