As VMware progresses towards virtualizing more and more business-critical and performance-critical applications, the tools and processes by which the virtualized data center needs to be managed need to change. As a first step, it is important to separate workloads into three groups: Tactical (no one cares if they go down for a while), Business-Critical (they must be up all of the time), and Performance-Critical (they not only must be up, but must deliver excellent response time all of the time).
VMware’s Virtualization Progress
At its VMworld conference in the Fall of 2012 VMware presented the following graph depicting its progress in virtualizing workloads to date and its expected progress. This graph shows that VMware is approaching virtualizing 60% of the servers in its customer base and expects to get to 73% in two years.
% of Workloads Virtualized (click image to expand)
Progress Virtualizing Business Critical Applications
Even more impressive than the fact that the overall virtualization penetration is approaching 60% is the fact that VMware has started to accelerate the progress in virtualizing business-critical applications. Notice that VMware has accelerated it progress in virtualizing Oracle databases from 3% year over year to 7% year over year and has accelerated the progress virtualizing SAP from 10% year over year to 12% year over year. It is fair to say that SAP is a business-critical application and that in most cases an Oracle database is part of a business-critical application, so this demonstrates tremendous progress with both the most important applications and those where the business owners have the greatest levels of concern regarding the impact of virtualization upon the operation of the applications.
% of Workload Instances that are Virtualized (click image to expand)
Operations Management and Infrastructure Performance Management for Business Critical Applications
If you are going to run business-critical applications on a virtualized infrastructure, you have to have an operations management solution in place. The objective of an operations management solution is to prevent resource -based and configuration-based issues from impacting the availability and performance of those business-critical applications.
Furthermore, if you are going to run performance-critical applications on a virtualized infrastructure, then they will need to deliver excellent response time to their users all of the time. In order for this to occur, the infrastructure will have to deliver excellent end-to-end latency to the workloads all of the time. You will not be able to deliver excellent latency to all of the workloads unless you measure it, which is what the Infrastructure Performance Management category in the table below is all about.
|Vendor||Product||What Makes Them Special|
|Astute Networks||ViSX||Simple to deploy network-attached Flash Storage optimized for VMware and deployable by the VMWare Adminstrator. Is based upon theÂ ViSX G4 High-Performance VM Storage Appliance.|
|Cirba||Cirba||Sophisticated predictive analytics for controlling capacity, resource allocations, and placements.|
|CloudPhysics||CloudPhysics||The CloudPhysics platform helps you make the best decisions in the face of continual change. We make the invisible visible – revealing otherwise hidden structures and relationships in your datacenter.|
|Dell/Quest (vKernel)||vOps Manager||Provides control over virtual environments and cloud deployments with analytics, advice, and automation. Includes performance impact modeling for environment changes, automated environment change implementation, VM issue diagnosis screen for all problems affecting a VM, VMware vCloud Director integration.|
|Hotlink||SuperVISOR||Extends VMware vCenter to fully manage Hyper-V, KVM, XenServer, CloudStack, and Amazon EC2 â€“ no additional console required.|
|HP||Virtualization Performance Viewer||Virtualization Performance Viewer (vPV) is a real-time diagnostic and triage tool that is lightweight and provides quick time-to-value. Administrators can visualize data from a single console in minutes and begin to diagnose performance issues rapidly using HP vPVâ€™s unique viewing capabilities.|
|Intigua||Â Intigua||Virtualizes your Management Agents so that their impact is minimized and so that they are easily managed at scale.|
|ManageEngine||Â IT360||Agentless monitoring for applications and servers along with single pane of glass view across physical, virtual, and cloud IT infrastructure. Includes privileged password management for VMware vSphere, capacity planning reports and automated resource optimization for virtualization software.|
|PHD Virtual||Â VirtualÂ Monitor||End-to-end virtual infrastructure monitoring for vSphere and XenServer & XenDesktop environments.|
|Reflex Systems||Virtualization Management Center||Real-time, streaming data analysis including customizable analytics, charts and graphs that give enterprises extremely granular visibility, and pure data fidelity for managing virtual infrastructure.|
|SolarWinds||Virtualization Manager||SolarWinds Virtualization Manager delivers integrated VMware and Microsoft Hyper-V capacity planning, performance monitoring, VM sprawl control, configuration management, and chargeback automation; all in one awesomely affordable product thatâ€™s easy to download, deploy, and use.|
|Splunk||Splunk App for VMware||Unlock the value of vSphere data and get granular performance management and analysis, as well as security reporting and monitoring, complete operational visibility, and capacity analytics. Includes the new Splunk App for VMware.|
|Veeam||Â ONE||Veeam ONE is a single solution for powerful and easy-to-use monitoring and reporting for VMware and Hyper-V and for Veeam Backup & Replication. It provides complete visibility of the backup and virtual infrastructures and affordably delivers the capabilities that matter most to virtualization administrators.|
|Veeam||MP for SCOM||The Veeam MP leverages and protects the investments you’ve already made in System Center by bringing VMware into the fold. It preserves the integrity of your unified Operations Manager console and the IT policies and practices you’ve built around System Center. You maintain compliance even as your environment changes and grows.|
|VMTurbo||Operations Manager||Ability to guarantee resource allocations based upon workload priorities. Includes integratedÂ configuration management with OpenStack|
|VMware||vCenter Operations||Automated operations management solution with self-learning analytics, application awareness and integrated performance, capacity and configuration management for virtual and cloud infrastructure. Includes full transaction tracing and closed loop integration from provisioning to monitoring to auto scaling.|
|Zenoss||Service Dynamics||A single product that delivers end-to-end service assurance for real-world, hybrid IT that spans physical, virtual, and cloud-based infrastructure.|
|Â Infrastructure Performance Management|
|Dell/Quest||vFoglight Storage||vFoglight Storage monitors virtual environments and provides performance and capacity metrics to help you better manage physical storage for virtual infrastructureÂ performance. vFoglight Storage supportsÂ EMC CLARiiON (CX3, CX4),Â HP EVA,Â HDS AMSÂ Filers, andÂ NetApp Filers.|
|Gigamon||GigaVUE Fabric Manager||Gigamon empowers infrastructure architects, managers, and operators with visibility into the physical and virtual networks without affecting the performance or stability of the production environment. Is based upon theÂ GigaVUE-FM platform to manage distributed GigaVUE-VM virtual fabric nodes on virtual and physical infrastructures.|
|Virtual Instruments||VirtualWisdom||Virtual Wisdom analyzes every frame header on a fibre channel SAN at line rates of up to 8 Gbits per second. It detects application performance slowdowns and transmission errors by measuring every SCSI I/O transaction from start to finish, for every server/volume combination (initiator/ target/ LUN). It is the only true real-time SAN performance monitoring device in the industry.|
|VMware||vC Ops for VNX||The EMC VNX Storage Analytics Suite enables storage operations management with customizable, intuitive dashboards. It will support block and file storage as well as displayEMC Fully Automated Storage Tiering (FASTâ„˘) Cache configuration and performance, and FAST VP (FAST for Virtual Pools) tiers and policies.|
|Xangati||Xangati Management Dashboard (XMD) Suite||Live and continuous tracking of infrastructure performance health (server+hypervisor, storage and network). Problems identified within first hour of install. IncludesÂ StormTracker: Industryâ€™s first solution to track and anticipate transient, performance storms that fly under the radar of conventional monitoring solutions.|
How to Pick a Solution
The most important thing to understand about the table above is that although vendors and products are grouped into two groups, there are vast differences in functionality and benefits across members of the groups. We are in the process of updating our white paper which will go through all of this in great detail, but some important considerations are:
- Everyone in the table above except Astute Networks, Hotlink, and Intigua can be broadly considered to be monitoring solutions. Astute Networks is about improving the storage performance of the environment by inserting flash appliances into the storage fabric. Hotlink is about allowing you to use vCenter to manage multiple virtualization environments and public clouds and to seamlessly migrate workloads between all managed environments. Intigua is about applying application virtualization techniques to the management agents in your environment so that they can be easily managed, updated, and controlled.
- The scale of the environment that the solution targets differs vastly across the solutions in table. CloudPhysics, Reflex Systems, Splunk, and Xangati all use big data architectures in their back ends to allow them to address very large virtualized data centers that need to be monitored with high fidelity. This was addressed in detail in Big Data Operations Management. Virtual Instruments has proven that it can measure end-to-end storage latency at the one-second level of granularity for the largest and most complex fibre channel SAN environments. On the other side of the spectrum, solutions like ManageEngine and SolarWinds are extremely easy and affordable to implement for SMB/SME scale environments that do not need high fidelity monitoring.
- Operations Management and Infrastructure Performance Management are two very different things, targeting two very different problems. Operations Management is essential as soon as you put business-critical workloads in your virtual data center. Infrastructure Performance Management is essential as soon as you put performance-critical workloads in your virtual data center. It would be entirely reasonable to purchase one of each of these in most cases.
Virtualizing business-critical applications requires an operations management solution that is designed from the group up for a dynamic and shared data center. Virtualizing performance-critical applications requires an infrastructure performance management solution that can monitor the end-to-end latency of the infrastructure in real time or near real time. As enterprises virtualize more than low-hanging fruit, these solutions will become essential elements of the virtualization process.