In VMware and the Ionix Assets – A Deeper Look, we took a fairly in depth look at the four products that VMware bought from EMC, and posited that VMware was now well on its way to fulfilling its promised intentions of becoming a vendor of a management stack for virtualization. In this article we take a look at the impact of these acquisitions upon the virtualization performance management market and the ecosystem of solutions available in this market.
This is one of the most important aspects of the overall topic of virtualization management for one very simple reason. VMware’s customers have virtualized the low hanging fruit that was and is under the direct control of the IT department. What is left to virtualize is servers that run applications that are business critical and performance critical. These applications have very powerful and vocal teams that own and support the applications and equally powerful and vocal constituents that use these applications.
Virtualizing these applications changes many aspects of how they are managed. Some of these changes include:
- The ability for the teams that own the applications to throw money at hardware and over-provision is either taken away or dramatically reduced as the financial motivation to virtualize is in fact to drive up the utilization of hardware resources and to save money in the process.
- Since IT owns the virtualization layer which will now live in between the operating systems and the hardware, IT is now more responsible for the performance of the applications then IT was when the application ran on physical hardware.
- Since virtualization will be the new layer in the stack, it and the team that owns it (IT) will be guilty until proven innocent.
- In the physical server and network environment infrastructure and applications performance were frequently approximated by inferring performance (response times) from resource utilization statistics. This no longer works for virtualized systems and is covered in detail in our Virtualization Performance and Capacity Management White Paper. The fact that this no longer works breaks most of the infrastructure performance and applications performance management products that have been used for years on physical infrastructure.
- In addition to breaking most of the legacy performance management products, virtualization also introduces a set of new requirements that have to be met in order to manage either the performance of the infrastructure or the performance of applications running on a virtual infrastructure. This is again covered in detail in the Virtualization Performance and Capacity Management White Paper.
The above factors have created a new market opportunity for how one monitors resource and availability, infrastructure performance, and applications performance in virtualized environments. A significant ecosystem of vendors has sprung up to address these opportunities. The balance of this article looks at how the acquisition of the Ionix assets by VMware impacts these markets and the available vendor solutions.
Application Discovery Manager
Of the four products that VMware bought (Server Configuration Manager, Application Discovery Manager, Service Manager, and Application Stack Manager) only Application Discovery Manager has a direct impact upon the Performance Management space. Configuration Manger and Service Manager address aspects of Virtualization Management and Application Stack Manager reduces the memory used by popular server based operating system and middleware. Application Discovery Manager consists of an appliance that does deep packet inspection of the network through a spanned or mirror port to indentify applications signatures in protocol stream. Application Discovery Manager (when integrated with VMware which should take between 12 and 24 months), will allow VMware to have a standard way of indentifying applications and their topologies when those applications run in the VMware environment. Application Discovery Manager is currently also not limited to discovering applications in a virtual environment. It can discover those applications in a physical environment as well, and since EMC Ionix will continue to resell this product, one would presume that VMware will not cripple it and take the physical discovery capabilities away.
Impact upon the Virtualization Performance Management Ecosystem
The presence of Application Discovery Manager in the VMware portfolio of products could either be very good for the third party ecosystem, or could represent a small but manageable problem. The good could occur if VMware makes Discovery Manager into a platform with API’s as well as a product. This would allow third party vendors to hook into it for the discovery of the software and hardware assets and configuration in the environment, and allow the ecosystem to concentrate on collecting unique data and providing unique root cause analysis and other valuable features. If Discovery Manager does not end up being open for the ecosystem, then vendors will have to continue do what they are doing now which is to build their own discovery into their products.
The diagram below depicts the three distinct segments of the Virtualization Performance Management ecosystem. Resource and Availability Management focuses upon collecting, storing, trending, reporting and alerting on resource utilization data – primarily for the purpose of capacity management and planning. Infrastructure Performance Management is a new category that has emerged that focuses upon determining Infrastructure Response Time, the key metric by which IT Operations can know of the infrastructure is performing as required for the application. Applications Performance Management deals with performance (response time) of key business applications, and also often provides deep dive and root cause analytics for troubleshooting.
The Resource and Availability Monitoring Space
It turns out that the biggest impact upon the market for resource and availability monitoring solutions by VMware does not come from this acquisition. VMware has an internally developed solution, CapacityIQ which is targeted at capacity planning and management through an analysis of how resources are used and growing in the VMware environment. VMware also picked up Hyperic through the SpringSource acquisition, which is a very robust and scalable resource and availability monitoring solution with broad support for many different types of servers, middleware products and networks. Therefore the case can clearly be made that even without the Ionix assets, VMware already had the assets to field a competitive offering in this space. Of course VMware has a significant integration and rationalization effort to go through, but it should be an entirely reasonable expectation for VMware to field robust and competitive offerings in this space by mid-2011.
The Infrastructure Performance Management Space
Infrastructure Performance Management is distinguished from the above category in that it does not focus upon collection of resource utilization statistics as a proxy for determining the performance of the virtualized infrastructure. Rather these solutions perform the essential function of directly measuring how long it takes for the infrastructure to respond to requests for work. The key metric here is Infrastructure Response Time. CA/NetQos and Xangati do a great job of getting a deep view of how the network (LAN and WAN) is impacting IRT and in fact use the network as a measurement point to determine their view of IRT. Virtual Instruments is the only vendor that can provide a comprehensive, deterministic, real-time and continuous view (all of the data all of the time) of the SAN. Akorri is the only vendor that can provide and end-to-end view of IRT, starting the instrumentation process in the guests and capturing the response time of every layer down through the physical spindles in the arrays of the storage devices. The acquisition of the Ionix assets has no impact upon this category as VMware did not have an entry in this category before these acquisitions and does not have an entry after these acquisitions.
The Applications Performance Management Space
The acquisition of Application Discovery Manager puts VMware in the position of having a strong application discovery capability, which is a key capability for applications performance management of a virtual environment. ADM probably has a breadth of discovery capabilities that is not matching in any of the APM products from the third party vendors in this space. However, several of these vendors have discovery of transaction flows between VM’s in a resource pool, and even discovery of transaction flows between VM’s in internal and external clouds (see Cloud Applications Performance Management gets Serious). The existing VMware AppSpeed product is a fine APM solution for Web/J2EE/.Net/SQL/Oracle oriented applications running on VSphere. However customers looking for a virtualization competent and cloud competent APM solution should consider the value of looking at solutions that are not tied to one virtualization platform, as it is likely that applications will span physical and potentially multiple virtual platforms in the future. Vendors like AppDynamics, BlueStripe, Coradiant, New Relic and OPNET all have very competent APM solutions that also have the architecture and feature set to function effectively in a virtualized environment. The acquisition of ADM does nothing to impact the position that these vendors have in this space. Finally there is the question of measuring the true end user experience from the perspective of the user’s workstation. This the area where Aternity and Knoa specialize, and there is not even a hint of a competing offering from VMware.
Existing VMware offerings competed in the Resource and Availability Management space prior to the acquisition of the Ionix assets, and the acquisition has done nothing to change the fact that vendors in this space face strong competition from VMware (or certainly will do so once Hyperic is integrated and ships as a VMware product). Infrastructure Performance Management is the key category that IT Operations needs to focus upon to understand the performance of their virtual environment, and the acquisitions do not change the positions of Akorri, CA/NetQos, Virtual Instruments and Xangati in this space. Adding ADM to VMware’s assets in the APM space adds a significant capability, but at the end of the day does not yet put VMware in the position to be able to provide an APM solution across physical and potentially multiple virtual environments as can AppDynamics, BlueStripe, Coradiant, New Relic and OPNET.
Share this Article:
Latest posts by Bernd Harzog (see all)
- VMware vSphere 6 Attacks Amazon with “One Cloud, Any Application” - February 9, 2015
- VMware vSphere 6 Attacks Red Hat: VMware Integrated OpenStack - February 3, 2015
- Will the Public Cloud Be the Next Legacy Platform? - January 20, 2015