Once reigning supreme in IT infrastructure tool chests, traditional data centre management tools are now facing a threat from smarter and more efficient virtualisation host tools. Can IT professionals rescue these data centre management tools?
As virtualisation in the data centre matured, IT pros turned to
Continue Reading This Article
Enjoy this article as well as all of our content, including E-Guides, news, tips and more.
How virtualisation is wreaking traditional data centre management tools
In many cases, an IT infrastructure’s data centre management tools may still appear to work, but with the virtualisation layer in place, they may be providing inaccurate data to the extent that they can't always be trusted.
Here are four ways that virtualisation may be breaking management tools:
- Operating system to physical host mapping
With a hypervisor layer in the IT infrastructure, there are now multiple virtual machines (VMs) on each physical server. Multiple VMs represent multiple operating systems and multiple applications. In many cases, physical hosts can run more than 100 VMs.
While this improves IT efficiency and speed from a business point of view, it makes data centre administration tasks a bit tricky. Traditional data centre management tools were designed under the pretext that for every physical host, there was one operating system (along with its applications). And every physical host would have just one hostname and just one IP address.
With virtualisation, no longer is there 1:1 mapping of operating system to physical host, and traditional management tools weren’t designed to take that into account. As a consequence, the results data centre pros get from these tools could be flawed.
- Constant change
Thanks to virtualisation’s hardware independence and virtual machine portability, VMs can be easily moved from one host to another. While IT pros could manually move a VM, they can do it automatically and quickly using hypervisor tools such as VMware vMotion or VMware Distributed Resource Scheduler (DRS).
With DRS, VMs could be moving around all day long without requiring IT’s attention. High availability (HA) features such as vSphere HA automatically restart VMs from failed hosts on other hosts (again, without IT’s intervention). In addition to moving VMs, the virtual storage of a VM (the virtual disk) could be moved around automatically. With VMware’s storage DRS (SDRS), virtual machine disk files could move from data store to data store automatically, without any effort from IT pros.
Traditional data centre management tools just weren’t designed with this level of change in mind. They were designed for operating systems and storage to stay put on each host. If something changed in the environment, then IT admins would have to manually update those changes in the management applications.
- No longer the most efficient way to measure performance
In many cases, data centre management tools just don’t work as efficiently as they should when virtualisation is put in place. For example, traditional management tools might contact every operating system using SNMP (or similar tools) to pull performance stats. But in a virtualised infrastructure, the hypervisor management tool can provide data on the performance on a per-VM basis.
Traditional applications such as backup and recovery tools also need to evolve because they run agents on each physical server to gather the data. This is not required in a virtual infrastructure where backup tools can contact the virtualisation host to get just the changes for virtual disks, making the backup process much faster and more efficient.
The role of data centre management tools within change management functions is also becoming redundant. The virtualisation host and virtualisation centralised management tool keeps a record of all those changes for all VMs, so data centre tools no longer need to go to each and every OS.
- Traditional data centre tools not talking to the right tools within the infrastructure
In many cases, traditional management tools just “aren’t talking to the right people” in the data centre to extract the relevant information that IT pros want -- making them inefficient and incorrect.
There was a time when data centre management tools were the “heroes” of a virtual infrastructure. Today, the heroes are virtualisation hosts and centralised management hosts. They report on everything when it comes to virtual machines including where they are running, performance data and more.
The fix is in
IT pros and management tools vendors can rescue the dying data centre tools. Here’s how these tools should be updated to become relevant in the virtualisation and cloud era:
- They must understand the hypervisor layer and one-to-many ratio: Data centre management tools recognise that there is a new layer in the data centre -- the hypervisor layer, otherwise known as the “virtualisation layer”.
This layer must be represented in the tools, and the tools need to represent this layer in their diagrams and reports. Tools should show whether the physical server has a hypervisor and how many VMs are running on it.
- They must talk to the virtualisation host or centralised tool: Tools need to be able to communicate with hypervisors through API calls. Only by talking to the hypervisor will tools know which VMs are on that host, the performance of the hosts (and VMs). Many virtualisation vendors publish APIs that help management tools record hypervisor data.
Such communication will help these tools provide other valuable information to IT pros such as the performance data of hosts, clusters, and VMs, what virtual disk blocks have changed, and much more.
Additionally, data centre alerting and monitoring tools may not need to “ping” or check on each VM anymore because a hypervisor knows whether or not the OS inside a VM is responding.
- They must work in real time to update changes: Management tools must be in constant communication with virtualisation hosts and centralised tools to get the latest information. A “nightly poll” – for which traditional data centre tools are used -- isn’t adequate in the cloud era.
- Use virtualisation by deploying as an appliance and using a plug-in: IT shouldn’t need a physical server for management tools in the data centre. These tools should be able to support and embrace virtualisation. The best way vendors and IT can enable this is by making the tools downloadable as a virtual machine.
Ideally, it would be an open source virtual appliance so that IT does not have to purchase or install Windows OS. In case of a VMware infrastructure, vendors could offer a downloadable “vApp” for easy imports into VMware vSphere or even Workstation and Fusion (for evaluation). All of this makes the evaluation and deployment simple and is the smart way to take advantage of virtualisation.
Tool vendors should create virtualisation platform client plug-ins that allow IT pros to access the tool within the virtualisation client. If vendors can’t do that, they should at least have a Web front-end that works from any computer on the network. Most admins no longer desire traditional Windows-App GUI front-ends.
How cloud computing is affecting data centre management tools
Infrastructure as a Service (IaaS) -- facilitated by cloud computing -- is an extension of virtualisation. The IaaS cloud layer rides on top of virtualisation and provides features such as orchestration/automation, self-service, catalog of applications and resource controls for the business units of the organisation.
To adapt to the IaaS cloud, management tools will need to:
- Understand how the downloadable vApp can play into the cloud. The vApp package can provide performance reporting, backup and other tool functions on a “per-vApp” basis.
- Understand the virtual data centre. Virtual data centres are just like physical data centres but in virtualised forms. Each organisation using an infrastructure cloud has its own organisational virtual data centre. The cloud provider could create multiple physical virtual data centres.
- Talk to the cloud director. Just as tools must talk to the centralised virtualisation management tool, they also will have to communicate with the cloud controller using published APIs. A cloud controller really knows about what VMs, vApps, and other cloud units and data centre management tools must be able to bridge this information to data centre admins.
Data centre management tools must adapt to the new virtualised and cloud-based data centres by understanding that servers have been consolidated in the virtualisation layer if they want to remain relevant and useful in the cloud era.
David Davis is the author of the VMware vSphere video training library from TrainSignal. He is a vExpert, VCP, VCAP-DCA and CCIE #9369 with more than 18 years of enterprise IT experience. Davis is a regular contributor to SearchVirtualDataCentre.co.UK.
This was first published in August 2012