| |
In the traditional IT stack there was a rigid 1-1 mapping between hardware, an instance of an operating system and a single software application. That rigid model lead to tremendous under-utilization of hardware resources. The industry statistic is that in this traditional model, servers are utilized only 5-15%. This is a huge problem for companies – having a very large pool of resources that stays idle most of the time.
But the story doesn't end there – the server sprawl and the associated under-utilization of resources have ripple through effects for the entire environment – server sprawl means not only wasted investment in hardware, but also unsustainable power, cooling, and real estate costs. This tremendous complexity means that it is hard to provision new infrastructure and to respond to changing business needs. IT departments are stuck wasting cycles on mundane tasks, and don’t have time to focus on what really matters. For example, in most companies a single sysadmin can support only up to 20 servers, and the time for provisioning a new server is often 6-8 weeks.
| | | Virtualization breaks the rigid bond between hardware and software, and allows multiple instances of an operating systems and multiple software applications to run on the same hardware box.
In essence, virtualization “bundles” or encapsulates the operating system and a software application into a virtual machine. This entire “package” of virtual hardware CPU, memory and networking, OS and application, is turned into a single software file. Virtual machines are hardware independent, and because they are files, they can be manipulated with the ease of file copy and paste. Virtual machines bring an entirely new level of efficiency and flexibility to the IT environment.
One way to understand the impact of virtualization on the IT industry is to compare it with electronic banking – once the cash money was turned into bytes, it increased the velocity of commerce – money can be moved around the world now with lightening speed – because it is virtual, it is information. By turning physical capabilities into information, virtualization makes provisioning and managing the entire IT infrastructure a lot faster and more flexible.
| | | How is virtualization typically used?
Infrastructure & Capacity Optimization (aka Server Consolidation) was the original use case for virtualization, and remains one of the biggest drivers for adoption. Deploying multiple workloads on the same server means that companies can do more with less – less servers, less storage, less networking cables, less space in the datacenter, less power and cooling. This use case is applicable to the overwhelming majority of servers in the datacenter.
But increasingly companies find that virtualization represents a new way of thinking about business continuity in all its forms – ensuring high availability of critical applications as well as planning for a disaster or recovering from a disaster.
A new, and very exciting extension of virtualization beyond the datacenter is what we call Virtual Desktop Infrastructure – the capability to host end user desktops in the datacenter, and expose them to end users through a thin terminal. This allows companies to drastically decrease the cost of end-user desktops, and apply the same rigorous security and data protection available in the datacenter to end-user desktops.
| | | Now let’s take a very specific look comparing the provisioning activity in the virtual versus physical world.
In the physical world, you have to procure and configure the hardware. This is an activity that can takes weeks if you need to order a new system. Then you go through the process of configuring the OS, networking, and applications. All together, this could be 20-40 hours of actual work, plus 4-6 weeks of time from start to finish.
By contrast, deploying a VM from a template takes minutes. Including configuration, you’re typically talking about <1 hour of actual work. Integrating into the typical IT procurement and approval process, and most likely you’re still only talking about 1 or 2 days.
| | | VMware’s hypervisor for the data center is called “VMware Server” First launched in 2001, we are now on the third generation, serving over 20,000 enterprise customers.
VMware Server is a bare metal virtualization platform, that is, it sits on top of the server hardware directly and allows partitioning of the underlying CPU, memory, network and storage resources between virtual machines that sit on top of it.
This hypervisor layer is an important choice to make especially because when you consolidate several servers onto a single one with many virtual machines, the hypervisor layer needs to be sturdy and reliable.
What are some of the key things to look for in a hypervisor?
Performance: Bare metal hypervisor architecture near native performance. VMware Server allows consolidation ratios a minimum of 10:1
Stability: >3 years of uptime at a customer, 86% of customers use in production, 43% standardize on VMs
Scalability: The hypervisor needs to scale to meet workload requirements with as many as 4 CPUs assigned to them,
Cross-platform support: HW agnostic, widest guest OS support including windows (multi-generations), Linux, Solaris x86, Novell
| | |
This makes it extremely easy to get started with virtualization. It’s as easy as:
- Plug in your server and boot directly into the hypervisor
- Do some minimal configuration
- Manage the server remotely and start running VMs within minutes
| | | The hypervisor then partitions a server into multiple virtual machines. VMs are basically just files and can be treated like files. This gives VMs amazing properties that then enable a much broader set of capabilities beyond just running many VMs on one server, not possible in the physical world.
These properties:
- Partitioning: With VMware Server you are able to run many different OS-es on the same machine
- Isolation: These VMs running on a physical machine are independent and unaware of each other. Infecting one doesn’t mean affecting all the others.
- Encapsulation: Since the entire OS + app is packaged up into a set of files – manipulation of VMs becomes very much simpler than their physical counterparts
- Hardware independence: Virtual machines can be created on any x86 hardware and moved to any other x86 hardware; in many cases live
| | | | | | | | |
For large scale consolidation projects, we recommend the use VMware Converter to automate the process of converting physical machines to virtual machines. Windows based machines can be hot-cloned ( i.e. a snapshot is taken of the running physical machine and used to create a virtual machine)
Converter can also be used to convert virtual machines between different types of VMware virtualization products – so if you’ve started with Workstation or Server, you can move your VMs to an ESX environment using Converter.
Third party images taken with Norton Ghost 9 or Symantec LiveState Recovery can also be converted to VMware VMs with this utility.
To conclude, virtualization is having a dramatic impact on how IT infrastructure is managed. Virtualization promises to:
- Increase the return on your investment in server technology
- Reduce the operational costs associated with managing and provisioning large and growing farms of servers
- Improve IT’s relationship and responsiveness to its business partners
- Increase uptime and improve server levels across 100% of your applications all while providing a rapid Disaster Recovery method.
|
|
