The Importance of Choosing the Right Workload Migration Tools Workload migration tools lay the foundation for projects like data center transformation and modernization and server consolidation. While there are many tools available to move a server workload from point A to point B, there are dramatic differences in features, performance, automation, customization and configurability. Administrators and project managers should read this white paper to understand how the differences in critical capabilities can affect overall project timelines and costs.
WHITE PAPER Table of Contents How to Do More with Less in Your Data Center................................................................1 Six Steps of a Successful Migration............................................................................2 Step 1: Inventory Workloads...2 Step 2: Assess (Planning and Analysis).....................................................................2 Step 3: Pre-migration...3 Step 4: Migration...4 Step 5: Post-migration...5 Step 6: Testing and Troubleshooting.......................................................................5 The Right Tools for Data Center Transformation...6 About NetIQ...7 WHITE PAPER: The Importance of Choosing the Right Workload Migration Tools
How to Do More with Less in Your Data Center It s a fact of life today that IT management is feeling pressured. Data center budgets are decreasing, even as infrastructure performance and availability demands increase. In an effort to help data center managers do more with less, Gartner recently issued a list of recommendations for data center modernization to CIOs, IT leaders and technical professionals. The common themes among the recommendations were: Planning and analysis Conduct an in-depth assessment of your infrastructure. Cost savings Determine how and where your organization can achieve savings in its server, storage and network infrastructure. Speed, automation and time-to-value Focus less on technical details and more on turnkey solutions that deliver speed and agility. Standardization Standardize as much as possible. Reduced complexity improves speed, agility and availability. Worry less about lock-in and more about ways to accelerate orchestration and automation of IT tasks. Administration costs (not hardware or software costs) are often the biggest contributor to escalating IT operations costs. These are all sound recommendations. But IT decision-makers aren t waiting idly for advice from Gartner or anyone else; they are busy implementing new data center strategies or overhauling the strategies they already have in place. This means migrating to higher-performance, energy-efficient servers, consolidating physical servers by leveraging virtualization technology to save on physical space and operational costs, and relocating and consolidating data centers in order to take advantage of cost savings from centralization and regional economics. One of the biggest costs of workload migration projects, and one that is often overlooked, is inefficient use of talent and resources. If IT personnel are manually completing migration-related tasks such as virtualization and server consolidation tasks that could easily be automated then they are unnecessarily taking resources away from other important initiatives. Instead of tying up valuable professional resources in an effort to capture greater operational efficiencies, a more practical course of action is to equip data center personnel with better tools. This paper takes a brief look at workload planning and migration tools, and differentiates NetIQ PlateSpin products as category leaders that offer the best overall value. We believe that this latter point best overall value is especially important since many competitive products are offered for free. Free tools seem like a better value up front, but they generally lack key features, speed and automation deficiencies that lead to longer, more expensive projects and poorer overall results. Decision-makers considering these freeware products should therefore consider how expensive free can really be. This paper illustrates that PlateSpin products are the most cost-effective even compared with freeware products and that they automate the processes of planning, migration and management better than any other products. IT decision-makers should be wary of too-good-to-be-true product offers that lack automation capabilities and that, in terms of total cost of ownership, end up being far more expensive than PlateSpin products, even when the initial cost of acquisition is zero. The merits of PlateSpin Recon and PlateSpin Migrate are highlighted in the following overview of the migration process. WHITE PAPER: The Importance of Choosing the Right Workload Migration Tools 1
Six Steps of a Successful Migration Step 1: Inventory Workloads Best Practice: List all servers, workloads and applications, as well as who owns what. This first step in any successful migration is proper planning. Its requirements are straightforward: Inventory the entire environment, including the potential source and target servers. Determine who owns workloads and servers. Establish what (if any) additional resources may be needed. Once this information has been gathered, you can filter out some portion of the server pool due to location, application owners, older or unsupported operating systems, and any other criteria that indicate unsuitability for migration. PlateSpin Recon Inventory Edition is a free standalone product that makes taking inventory of workloads and resources a simple task by identifying all the hardware and application services running in the data center. PlateSpin Recon Enterprise Edition also includes this functionality. As for determining who owns what and gaining their approval for the migration, PlateSpin Recon can expedite the process by labeling every server and associating them with a department, owner and application. This enables data center personnel to bundle scheduling announcements and requests rather than sending out 10 or 20 separate requests to the same user. Step 2: Assess (Planning and Analysis) Track resource utilization of all workloads over a 30-day cycle to deliver maximum virtual machine density and highest consolidation ratios. Record utilization based on usage peaks or averages spanning 24-hour periods instead of single peak points. Eliminate waste and inefficiency by discovering allocated but unused or under-used storage, disk, CPU and network resources within your existing virtualization environment. Be sure your assessment plan accounts for different CPU architectures. Clock speed and core count alone are not enough; advances in CPU architecture mean that a current-generation, multi-threaded CPU core may be two or three times faster (or more!) than an older, single-threaded core, even at the same clock speed. Compare different consolidation scenarios such as scale-up versus scale-out, consider different server makes and models, and identify cost savings from power and cooling. This step is all about measuring twice and cutting once, as the carpenter s adage goes. In other words, proper planning and analysis are the keys to minimizing the risks, costs and aggravation associated with any project especially one as critical as changes to data center infrastructure as well as shortening the overall project timeline. Workload migration projects can be much more complex than you might imagine, and only thorough and intelligent planning up front can identify the potential pitfalls and define a clear path to success. Planning at this level requires tracking resource utilization of all the workloads you might be considering. NetIQ recommends tracking workload resource utilization for a 30-day business cycle, with peaks or averages recorded in 24-hour periods. PlateSpin Recon Enterprise continually tracks and records workload resource utilization to maximize virtual machine density and consolidation ratios, while preventing performance-crippling resource contention. PlateSpin Recon Enterprise accomplishes this by matching the utilization peaks of some workloads with the utilization valleys of others. WHITE PAPER: The Importance of Choosing the Right Workload Migration Tools 2
PlateSpin Recon Enterprise also finds unused resources even when the virtualization infrastructure software shows 100 percent allocation. PlateSpin Recon Resource Reclamation reports identify identifies and reports gaps between resource allocation and actual utilization. Minimizing these gaps maximizes the virtual resource capacity of the existing infrastructure, extending useful service life and postponing new hardware purchases. PlateSpin Recon Enterprise analyzes the four critical dimensions of a workload CPU, disk (both total footprint and rate of change), memory and network over time, and even across thousands of servers simultaneously. As a result, it provides consolidation plans that maximize utilization while minimizing resource contention. In addition, PlateSpin Recon accounts for the performance characteristics of different CPU architectures. For example, unlike some free products, it doesn t rely solely on processor clock speed or number of cores. Instead, the PlateSpin Recon Capacity Planning Module recognizes the performance differences between a 3.0 GHz quad-core Intel Xeon processor and a 3.0 GHz quad-core AMD Opteron processor, thus increasing the accuracy of consolidation plans. PlateSpin Recon even compares and contrasts potential power and cooling cost savings to provide the return on investment (ROI) derived from various consolidation scenarios. Custom fields allow power and cooling requirements for major hardware platforms to be entered and maintained in a central database, enabling organizations to analyze and cost-justify green computing initiatives. In fact, PlateSpin Recon Enterprise enables you to create multiple custom scenarios with user-defined target server specifications, including server templates or existing virtual machine servers, to create an optimal consolidation plan. It can even help you improve the performance of your workloads by identifying hidden configuration errors in virtual machines. Once the planning is complete, PlateSpin Recon saves time by letting you export your chosen consolidation plan directly into PlateSpin Migrate, so you can seamlessly transition projects from analysis to implementation. PlateSpin Recon then exports all of the upfront planning and placement information, and everything is prepopulated in the PlateSpin Migrate console. And with PlateSpin Recon Enterprise, you can periodically reassess and reoptimize virtual infrastructure, reclaim and redeploy underused resources, and even forecast future requirements. Other workload migration solutions aren t nearly as comprehensive. VMware Capacity Planner, for instance, is a one-time planning tool that builds consolidation plans based on the busiest hour of the week and average utilization. Typically, free tools require manual adjustments to account for single- and multi-core processors, and those with or without multithreading. And without CPU normalization capabilities, other solutions tend to underestimate the true processing abilities of many modern CPUs (and overestimate the performance of older architectures with higher clock speeds), resulting in lower-than-anticipated utilization on the new virtual hosts. PlateSpin Recon also supports a broad range of virtualization platforms, making it the best solution for heterogeneous data centers. Whether you want to virtualize using VMware, Hyper-V, KVM or Xen hypervisors, and with Windows or Linux operating systems, PlateSpin customers can standardize using a single tool. Another planning issue for many enterprise customers today is that tools such as VMware Capacity Planner store and analyze infrastructure data offsite at the vendor facility. Because of privacy and security concerns, many organizations prefer to keep their sensitive data inside their own firewalls and under their own control. Unlike the competition, PlateSpin products are made for enterprises with high security requirements. The products are deployed and run in thousands of data centers by data center personnel, yet all data collected by PlateSpin Recon stays within the servers at the customer site. WHITE PAPER: The Importance of Choosing the Right Workload Migration Tools 3
Step 3: Pre-migration Configure migration jobs so they can run unattended. Schedule jobs to run during set downtime windows or when utilization is lowest. Take advantage of various options to configure (enable/disable) services on the target. The more you can do in this phase, the less time you spend in front of a console after the process is complete. PlateSpin Migrate not only automates the migration process, it also logs and updates the user on progress and allows you to schedule jobs to fit the environment not the other way around with or without the presence of staff. You can schedule workload migrations to run automatically in the middle of the night or on weekends. If a glitch occurs, disruptions are minimized and the event is recorded so that remedial action can be taken later; email alerts can also be sent out right away, if preferred. However, with PlateSpin solutions, production system disruptions are extremely rare because all testing and validation is done in the initial replication while the production system is still up and running it only requires scheduling a few minutes to do the final synchronization and cutover. In this way, PlateSpin Migrate minimizes or eliminates the prolonged outages typically associated with the migration process. While competitive products such as VMware Converter offer varying degrees of automation depending on the environment, the automation capabilities offered by PlateSpin Migrate are superior in every case. With VMware Converter, certain configurations require much more manual intervention, and regardless of the type of configuration, data center staff must be present during the entire migration. In addition, VMware Converter offers limited configurable options compared to PlateSpin Migrate. These shortcomings can often double or even triple the migration time. In addition to longer migration time, staff must man a console and manually configure the migration options instead of working on other tasks while an unattended migration runs automatically. PlateSpin Migrate not only automates the migration process, including post-migration configuration, but also allows job scheduling. Step 4: Migration Test workloads in the new location before going live. Consider staged migrations for migrations over slow WANs. When consolidating servers, relocating a data center, or undertaking any type of data center transformation project, you want to avoid unexpected results that create delays and can potentially force a complete restart. PlateSpin Migrate Server Sync is designed to help you avoid unexpected results. Server Sync lets you perform an initial transfer to the target site or host, test the workload in the new location while continuing to run the source, and then perform a final sync before cutting over the workload. Server Sync allows customers to test replicated workloads in their new environment, then synchronize only the changes before going live. You can repeat this test-and-sync step as often as necessary, or even stop the transfer completely and continue later from the last point. With competitive tools, administrators must pick one of two unappealing options: bring the source server offline and perform a second full migration, or omit testing altogether. PlateSpin Migrate also supports multiple concurrent migrations up to 40 which dramatically shortens both the migration timeline and overall time-to-value. VMware Converter and other free products perform well on individual migrations, but performance can degrade significantly as the migration count increases, especially to the same host. While competitors may boast claims of high performance based on laboratory conditions like moving a clean Windows install over an unused network connection to an idle, otherwise empty hypervisor this unrealistic, artificially-constructed migration scenario is nothing like any realistic production environment. PlateSpin Migrate is designed to provide superior performance over real-world networks, both local- and wide-area. WHITE PAPER: The Importance of Choosing the Right Workload Migration Tools 4
Staged migrations are often ideal, from both a timing and cost perspective, when migrating several servers over WANs a realistic scenario in many organizations. Staged migrations let you copy a workload to an image, physically ship the image to the required destination, copy it to the new location whether physical or virtual, and then synchronize only the changes over the network. With slow, expensive WAN connections, this is often the fastest and cheapest option. PlateSpin Migrate is especially useful for performing staged migrations over LANs or WANs, or to remote locations with limited bandwidth. In fact, PlateSpin Migrate was specifically developed to deliver faster file transfers over high-latency networks of varying quality. In the event of a network disruption, PlateSpin Migrate automatically resumes from the point of interruption when service is restored. Free solutions typically require perfect network conditions, and may abort migrations during even minor network interruptions, which costs your team hours or days when you are forced to restart the migration process. And even if the job doesn t abort, performance can be very poor over WANs and high-latency networks, or networks with high utilization. Step 5: Post-migration Use scripts to automate post-migration events to further decrease manual intervention and involvement. Stay flexible by retaining the ability to move workloads from anywhere, to anywhere. After the migration is finished, you may want to add additional software such as a virus scan utility. PlateSpin Migrate lets you write a script to include this post-migration step in some or all of your migration jobs. This level of customization of post-migration processes ensures initial time savings as well as ongoing savings by automating time-consuming, repeatable processes as part of the migration job. Scripting post-migration events before the migration even begins enables those events to take place automatically, so IT administrators don t need to be present during and after the migration. The only way to confirm the success of a project is through post-migration validation. Validation testing and the ability to take an iterative and ongoing approach to data center transformation can ensure the success of the initial project, while actually delivering on the promise of continued optimization and management. PlateSpin Migrate makes it easy to test and validate the success of a migration, without tying up valuable IT staff resources. You can configure PlateSpin Migrate to automatically generate a post-migration report that you can email to all stakeholders announcing the completion of the migration and showing performance gains and the end result. Detailed logs and reports can also verify compliance with government laws, industry regulations or trade group policies. In fact, PlateSpin Migrate has extensive job logging, audit trails and detailed diagnostics, and errors and warnings provide the information you need to identify, isolate and resolve issues quickly. In addition, the workload licensing and complete anywhere to anywhere (X2X) functionality offered by PlateSpin Migrate let users move workloads between physical servers, virtual hosts and image archives as needed. Other products are single-purpose tools designed to move workloads from a physical source to a specified virtual host. They are dedicated physical to virtual (P2V) tools that treat the virtual host as the final destination. With PlateSpin Migrate, you can move a workload back to its original location for testing purposes or to a new location, such as a newer virtual environment with better performance or lower costs, with no additional license requirements. WHITE PAPER: The Importance of Choosing the Right Workload Migration Tools 5
Step 6: Testing and Troubleshooting Take a strategic, multiphase approach to ensure that everything is running correctly at each level of abstraction. Prepare and run jobs in a way that minimizes the risk of failure. Testing is the part of the workload migration process that most often gets shortchanged. Infrastructure projects are expected to move at the speed of business, and this imperative to constantly speed things up can cause decision-makers to limit testing. However, the best way to minimize risk and verify the success of a project is to combine comprehensive planning with thorough testing both before and after migrations. This approach allows data center managers to look at data center transformations holistically, lays the groundwork for a one-time transformation and, most importantly, provides the basis for ongoing data center optimization. Whenever possible, testing should include multiple test stages that focus on the following business-level components of the data center: Individual server workload (the operating system, middleware, application and data stack residing within the server infrastructure) Application cluster (most applications are multi-tiered, running across multiple server workloads) Full data center (dependencies across applications, relied on for business services) A multiphase testing approach can ensure that everything is running correctly at each of these layers of abstraction. By testing the individual server workload components, the applications, and the data center as a whole, you can ensure the least amount of risk. The Right Tools for Data Center Transformation PlateSpin Recon and PlateSpin Migrate simplify and automate the data center transformation process with less risk and greater success than competitive products. In addition, unlike freeware and other one-off solutions, PlateSpin products retain their value to the enterprise by leveraging integrated planning and testing, and giving data center managers options that go far beyond legacy set-and-forget approaches, offering the foundation for continual, on-demand data center resource optimization. Whether you employ VMware, Hyper-V, KVM or Xen now, you ll be able to use PlateSpin Migrate and PlateSpin Recon as you explore other virtualization technologies. PlateSpin workload migration products offer broad multiplatform support for streaming workloads between any network hosts. This support extends to all of the leading virtualization technologies and hypervisors, as well as multiple operating systems and hardware configurations. PlateSpin Migrate also includes many additional features designed to make the migration process more secure (encryption and FIPS-compliant), faster (data compression), and robust under all network conditions (NAT support and bandwidth throttling). PlateSpin products provide organizations with mature, proven solutions for testing, migrating and rebalancing workloads across infrastructure boundaries from desktop to data center. They are the highest-performing, most mature and most cost-effective workload analysis and migration products on the market today. WHITE PAPER: The Importance of Choosing the Right Workload Migration Tools 6
About NetIQ NetIQ is a global enterprise software company that meets the demands of hybrid IT with solutions for identity and access management, security and data center management. Using our solutions, customers and partners can capitalize on the opportunities in today s complex and ever-changing IT landscape. By aligning technologies and service delivery methods, our customers can properly secure, manage and measure the many computing services operating across its physical, virtual and cloud computing environments. By doing this, they are better able to provide strategic value at the speed of business. Learn more about our award-winning software solutions at www.netiq.com. This document could include technical inaccuracies or typographical errors. Changes are periodically made to the information herein. These changes may be incorporated in new editions of this document. NetIQ Corporation may make improvements in or changes to the software described in this document at any time. Copyright 2015 NetIQ Corporation and its affiliates. All Rights Reserved. Access Manager, ActiveAudit, ActiveView, Aegis, AppManager, Change Administrator, Change Guardian, Cloud Manager, Compliance Suite, the cube logo design, Directory and Resource Administrator, Directory Security Administrator, Domain Migration Administrator, Exchange Administrator, File Security Administrator, Group Policy Administrator, Group Policy Guardian, Group Policy Suite, IntelliPolicy, Knowledge Scripts, NetConnect, NetIQ, the NetIQ logo, PlateSpin, PlateSpin Recon, Privileged User Manager, PSAudit, PSDetect, PSPasswordManager, PSSecure, Secure Configuration Manager, Security Administration Suite, Security Manager, Server Consolidator, VigilEnt, and Vivinet are trademarks or registered trademarks of NetIQ Corporation or its affiliates in the USA. All other company and product names mentioned are used only for identification purposes and may be trademarks or registered trademarks of their respective companies. Worldwide Headquarters For a complete list of our offices 515 Post Oak Blvd., Suite 1200 In North America, Europe, the Middle East Houston, Texas 77027 USA Africa, Asia-Pacific and Latin America, Worldwide: +713.548.1700 please visit www.netiq.com/contacts. U.S. / Canada Toll Free: 888.323.6768 info@netiq.com www.netiq.com www.netiq.com/communities WHITE PAPER: The Importance of Choosing the Right Workload Migration Tools 7