to commoditize or not?

After a few suggestions from the virtualization community to either drop Xen VMM for Hyper-V or to build tools on top of vmkernel let's hear yet another proposal

(1) acknowledging vrc's (http://www.virtualrealitycheck.net/) performance studies showing xenserver kernel outperforming vmkernel with xenapp workload and hypervisor scalability

(2) reinforming SAP's choice for XenServer easily doubling workload compared to physical machines (SAP XenServer/XenApp case study).


(3) saluting the growing number of partners in the Xen(Server) ecosystem: Intel, HP, Dell, NetApp, DataCore, Symantec, Vizioncore, Fujitsu-Siemens ...

(4) confirming Xen-kernel to be the foundation of XenServer, XenDesktop, XenClient, Xen Cloud C3, Hosted Xen et all

(5) embracing Xen communities development efforts


isn't it time for VMware to drop vmkernel in favor of xen-kernel?

(a) there were claims that binary translation (BT) is better than paravirtualization(PV). who has adopted vmi and paravirt_ops respectively

(b) We've seen corporate proof that PV performs better than BT -- pickup up here

(c) independent consultants prove better performance of Xen(Server) kernel

(d) Paul claimed they have 2.500 people in R&D. From a business perspective wouldn't it make sense to leverage the xen hypervisor development community and focus in house development on management tools?

Performance binary translations vs. paravirtualization

Richard McDougall (former DE at Sun, now Chief Performance Architect @VMware) published a detailed article on performance in VMware's corporate blogs.

Interesting to find: "In a recent study, the performance of Oracle 10g R2 using the Swingbench online transaction processing workload on a paravirtualized Linux guest shows a moderate gain of 10% when using paravirtualized CPU interfaces."