Now the VMware has release Capacity IQ it is worth taking a look at the category of Capacity Planning and Monitoring Tools for VMWare (and other virtualization platforms), and see how they compare to VMware’s offering. This article highlights a couple of the capabilities of each product and is not intended to be an exhaustive product review. More detail is contained in the White Paper available for download at the end of this article.
VMware has released CapacityIQ, its tool for helping VMware administrators manage and plan capacity for virtualized environments. CapacityIQ supports VirtualCenter 2.5, Update 4 and Update 5, managing hosts running ESX Server 3.0.2 through 3.5. CapacityIQ 1.0 does not support VMware vSphere 4.0 or vCenter 4.0. This is an unfortunate restriction (that we expect VMware to quickly address) as this tool would be very useful for organizations that are rapidly expanding their vSphere environments with Tier 1 applications and other business critical workloads.
That said, CapacityIQ collects and produces some extremely useful information for understanding the forward looking usage of key system resources. CapacityIQ includes an ability to simulate the impact of adding hosts to the cluster (the rise in the green line is capacity being added), and the impact of guests being added or subtracted from the cluster as well.
Akorri BalancePoint is unique in that it is the only product that collects all of the normal Virtual Center data, and also collects (and maps) I/O operations data down to the individual spindles in the storage arrays. Unlike solutions that simply report CPU and memory metrics BalancePoint also does a great job of “boiling up” this massive amount of data into highly useful comparative analytics. For example, BalancePoint calculates an Infrastructure Response Time metric on a per VMware Host basis that is the result of all of the queues and delays that exist between that Host and the actual spindles that the applications on that Host are mapped to. This IRT metric is then plotted against load to provide a Performance Index (PI) which is a highly accurate and easily understandable picture of where the Host is in terms the tradeoff between the amount of work being done and the IRT for that Host. BalancePoint calculates the optimal PI for your host, and and your current PI. This provides you with a very valuable tool for knowing how much further you can push the capacity utilization of your environment without either wasting capacity (and money) or causing performance problems.
BalancePoint can also calculate these metrics on a per application basis in each Guest, giving an accurate picture of how your infrastructure is performing for each of your key applications.
vKernel Capacity Analyzer
vKernel has been focused upon Capacity Planning and Capacity Management of VMware environments for longer than any other VMware focused ISV, and has an extremely mature solution with a significant customer base. vKernel collects key CPU, Memory, Disk Space, and I/O Activity metrics from Virtual Center, and automatically processes them through forecasting analytics that provide a forecast of when you are going to run out of each key resource by cluster or resource pool.
vKernel supports Virtual Center 2.0 and higher as its data source so it has good support for previously shipping as well as recently shipping (ESX 3.5) and new (vSphere) versions of VMware. vKernel does not support any virtualization platforms other than VMware.
Veeam Monitor is a great tool for other pieces CapacityIQ is lacking: actual performance monitoring and alerting, troubleshooting, as well as comprehensive trend analysis (especially with the latest release, which allows for trend analysis by business groups). The main benefits of Veeam Monitor are:
- Veeam Monitor offers performance management, troubleshooting and business-oriented trend analysis in one package for one price.
- Veeam Monitor is $250 per CPU socket, making it affordable for both SMB and enterprise customers.
- Veeam Monitor provides trend analysis from a technical as well as a business perspective, e.g. how storage consumption is growing in R&D vs. Marketing department in the company. It is integrated with Veeam Business View.
- Veeam Monitor supports all versions of VMware
- Veeam Monitor will support multi-hypervisor (VMware and Hyper-V) environments by adding Hyper-V support in H1, 2010.
BlueStripe FactFinder is not a capacity management product, rather it is an applications performance solution designed to work on applications hosted on physical and virtual infrastructures. BlueStripe is unique in that it automatically discovers application topologies and calculates hop-by-hop response times for every application (irrespective of how the application is architected).
FactFinder also collects the load on each tier of the application (and the application as a whole). Combining the load (Responses/Second) with Average Response Time creates a unique perspective on the capacity of environment hosting the application. You can make the argument that it really does not matter what the resource utilization profile of the infrastructure is as long as the applications are delivering the required level of performance to the users. FactFinder provides precisely this perspective.
Capacity Planning and Management of virtualized environments must be done with much greater frequency (at least daily) in virtual environments that in physical environments, and must also deal with challenges (outlines in the white paper referenced below) that are not present in physical environments. For these reasons enterprises pursuing capacity management in virtual environments are urged to take the following approaches:
- Clearly an understanding of how the infrastructure load is changing over time is needed in order to be able to have a load forecast, and to be able to know when to add physical server, network and storage capacity within the procurement windows for these resources. Therefore when it comes to managing the capacity of the virtualized environment, a good solution should capture the load being placed on all of the key physical resources in the environment (CPU, Memory, Network and Storage), and provide a good automatic forecast of the usage of these resources.
- However, just watching resource utilization is not enough. Knowing how intensely or heavily a resource is being used does not tell you what the impact of that usage is upon applications performance and end user experience. Therefore solutions that calculate infrastructure response time like Akorri BalancePoint should be given a strong look.
- While an understanding of infrastructure response time is essential to understanding how capacity is being used, infrastructure response time should be augmented with an understanding of how infrastructure usage is impacting applications performance. This is where BlueStripe FactFinder can make a very significant contribution to the overall issue of capacity and performance.