Better late than never. There were some really good (and not so good) technical reads this past month:
VMware vShield Zones – Reviewers Guide – If you haven’t figured this out yet, pay close attention to the Reviewers Guides. If you have even slightly thought of trying out a product or technology, the reviewers guide is the next best thing to having an engineer over your shoulder walking you thru the product. This is a really good one on Zones. I learned quite a few things about the product that I was unaware of. It’s a great read if you need to lock down and firewall off your VMs (or if you just want to learn how the VMs talk to each other).
Performance Brief for IBM WebSphere Application Server 7.0 with VMware ESX 4 on HP ProLiant DL380 G6 Servers – This is a very specific paper on running Websphere on HP servers. I did find some interesting bits in it however. Specifically, configuration tips to maximize performance running Websphere in a VM. The performance metrics should also be evaluated if you want to run Websphere on any hardware platform (perhaps even IBM).
Best Practices for Running vSphere on NFS Storage – I’m currently in a documentation war with my EMC Channel SE who swears by running VMware on NFS (You still have to convince me Steve, I’m a block-IO bigot). This is a must-have for all Admins running VMware on NFS. It’s also a really good guideline if you want to compare performance between file-level or block-level IO in your VMware environment and make sure your making an accurate decision. My favorite section of this paper: “Previously thought to be Best Practices.” Every best practice white paper should have that to debunk outdated information. Outstanding work VMware!
PVSCSI Storage Performance -This is a paper I was waiting to see. It compares the performance of the PVSCSI adapter to the LSI Logic adapter. I guessed pretty close on the outcomes. The PVSCSI adapter does perform better under higher IO workloads (some have stated only use it >2000 IOPs.) The only thing I didn’t like about the test was that RDMs were used. VMware has argued for the last few years that high-performance should not be a requirement for using RDMs. So why not use what the majority of customers use in their environments?
RIM BlackBerry Enterprise Server on VMware Virtual Infrastructure Deployment Guide – I really liked and hated this paper. I liked it because it has some really good best practices and deployment tips depending on the size of the environment. I hated it because it definitely contains errors: Table 2 shows a BES server with 23% utilization on ESX 3.5, Table 3 shows a BES server with 27% utilization under ESX 4, Figure 12 says the CPU load went down from 27 to 23 going to vSphere – not according to your tables kids.
That will do it for this month’s White Paper Review. We’ll see you next month and look for more great technical information.