Virtualization is a process, not a destination. If it were a destination, would it be called vNerd Island? vGeek Paradise?

A New Beginning. A Journey Into the Flourishing Paradise of Virtualization Nirvana.

This is my first blog in a new technology area that I’ve been very involved with over the past year. I’ve worked my way into the promised land of virtualization in the process of achieving the VMware Certified Professional (VCP) and VMware Certified Instructor (VCI). This is my first blog out of many future blogs in which my role will be that of virtualization enthusiast. I hope you enjoy the blog, ask questions, post your own personal virtualization experiences, and feel free to collaborate. I'll be talking about VMware a lot, but eventually we'll talk about the Cisco Nexus datacenter switch platform, and the Cisco Unfied Computing Server (UCS)platform as well. Virtualization is a process, not a destination. I suppose we can say the same for happiness, but I’ll try to not veer off into metaphysics. The virtualization movement normally starts with server virtualization. Server virtualization is an easy sell because it’s very easy to quantify a return on investment (ROI). The ROI in server virtualization is normally directly related to the physical to virtual (P2V) conversion process. vCenter Converter is a vCenter server component that enables the process of converting a physical server to a virtual server where a complex operating system with thousands of files and folders (directories) is converted to a handful of discrete files. We’ll get into the files eventually in this blog, but I’d like to first also mention that there is a free vCenter converter standalone application that you can download and convert your machine to a virtual machine right now. You can convert your Microsoft Windows desktop or laptop to a handful of files and run your Microsoft Operating System on your MAC with VMware Fusion. I’m writing this on my Dell laptop with Windows 7 SP1 right now, but I have a MAC book pro upstairs waiting for me to have some extra time to work on it. Anyway… back to virtualization in the Enterprise deployment context. Many virtual servers can run on one physical hardware server enabling anywhere up to 12:1 server consolidation ratios. Consolidation ratios will vary from that of 4:1 to maybe higher than 12:1, but 12:1 is on the high side. Virtualization candidates are driven by many factors, but the average CPU utilization of the physical server is a key component that should be taken under consideration. VMware’s Guided Consolidation tool monitors servers for activity and returns a virtualization confidence factor. This is a great tool to use before converting your physical server. Not all servers are good candidates for virtualization at this time and it’s much better to find this information out before you do the conversion… or you may have a VERY interesting week at the office… This career limiting event (CLE) could get you into the Donald Trump “You’re Fired!” conversation in some environments. High throughput databases and application servers with high network throughput requirements normally need the ability to write their information to the underlying server hardware faster. These types of servers are often not good candidates for virtualization. VMware (and the virtualization community) has been working with Intel and AMD to enable faster I/O technologies by enabling hardware I/O devices like network interface cards (NIC) and storage host bus adapters (HBA) the ability to write directly to I/O devices, bypassing the virtualization layer which normally performs binary translation. Binary translation is more time consuming than direct I/O technologies like paravirtualization SCSI drivers and VM Direct Path I/O technologies incorporated with vSphere 4. Paravirtualization drivers provide higher throughput and lower latency than virtual device drivers. Intel and AMD’s virtualization technologies (Intel VT and AMD-V) support paravirtualization in hardware in newer CPU models. OS assisted virtualization is an option in some Operating Systems (Windows Server 2003, 2008, Red Hat Enterprise Linux (RHEL) 5 and above), but hardware support in the processor (CPU) is faster than software assisted virtualization. AMD-V marketing has identified this capability as the Input/output memory management unit (IOMMU), while Intel VT updated their marketing to Intel VT-d to indicate that their chips fully support paravirtualization. VMDirectPath I/O is another technology that’s pushing the envelope in the direction of end to end datacenter virtualization. VMDirectPath I/O allows the guest operating system (virtual machine) to make direct calls to the hardware I/O devices in the physical server through the IOMMU to one of the four food groups (CPU, memory, storage I/O, network I/O). VM Direct Path I/O is experimentally supported at this time which means VMware does not support the use of the technology in production, but you can start using it in the lab. There are many technologies like VMotion that are not supported with VM Direct Path I/O. 10 Gigabit Ethernet (10GE) networking will benefit greatly from the deployment of VM Direct path I/O. VMware has noticed that the virtualization overhead of 10GE networking was overtaxing the processor. Continue with me on this journey and we’ll talk about ESX, ESXi free (re-branded to “VMware vSphere Hypervisor” in 4.1), ESXi embedded, ESXi installable, vSphere, vCenter server (formerly virtual infrastructure [VI] server), VDI, vMotion, Distributed Resource Scheduling (DRS), Distributed Power Management (DPM), Dynamic Voltage and Frequency Scaling (DVFS), High Availability (HA), Fault Tolerance (FT), and lots of other interesting three letter acronyms (TLA) that are not coming to mind right now. Sound like vNerd paradise?

Join the Network World communities on Facebook and LinkedIn to comment on topics that are top of mind.

Copyright © 2011 IDG Communications, Inc.

SD-WAN buyers guide: Key questions to ask vendors (and yourself)