Succeeding with Cloud Management Restore Order. Prevent Chaos.
Contents Introduction: It s Easy, But... 3 Commence Cloud Chaos... 3 Restore Order... 5 Your Next Move... 7 About egroup... 8 2014 by egroup Inc. All rights reserved. egroup Corporate Office 482 Wando Park Blvd. Mt. Pleasant, SC 29464 1.877.eGroup.1 (1.877.347.6871) www.egroup-us.com 2
Introduction: It s Easy, But Virtualization has become so easy that everyone from consumers to enterprise IT departments are making use of cloud based technology. In fact, products from companies like VMware make implementing the basic benefits of the cloud almost trivial. This ease of implementation, however, can often times be the kiss of death as little attention is paid to actually managing these clouds once they are stood up. So, as you contemplate your next course of action to transform from a useful, but unwieldy set of utilities, into a well-managed enterprise-level cloud, consider the recommendations and tips we offer in Cloud Management Done Right. Commence Cloud Chaos If you have the basics of your vsphere in production, your organization is already benefiting from the most mature x86 virtualization platform available. But as your environment evolves and expands to include more capacity and additional applications, you may be experiencing some of the issues that crop up because your deployments are not being managed in a holistic manner. Ensuring a management system is in place that can encompass the entire virtualized infrastructure can bring order to the impending chaos. For example, following a period of growth and use, issues inevitably arrive that you didn't expect. Here are a few that are all preventable with the right mix of management components: Time to deploy virtual machines - What was originally a dream come true now seems to take more time than you think it should. It's still quicker to set up a virtualized machine than to install new hardware 3
and configure an operating environment, but now that you and your company know what's possible, creating a new instance seems like it should take less time. Failure to track VM ownership and life cycles - Tracking who requested an instance, then finding whether they are still using it can mean messaging all possible users. Chances are they don't even remember which instance, how to identify it, or what they needed it for. The result is that you have a bunch of zombie VMs that never get retired and just get dragged along through all of the infrastructure upgrades. Hard to identify root causes of VM slowness when problems are reported Finding which virtual machine is causing issues, what those issues are, and eventually determining what the fix might be can take more time than moving the impacted users to a new instance on a different cluster. Falling into VM Creep is a trap that's easy to succumb to just because it's so simple to create a new virtual machine whenever needed. Too many VMs occupying a limited resource can cause significant performance issues. Of course, having the flexibility to add those VMs is also one of the main benefits of the technology. But all things need to be taken in proper measure. The problem is that it's difficult to determine how many is too many. Every combination is unique and can consume different resources at different times. Working around performance issues can boil down to adding more hardware to increase the resource pool, but that's likely to be a never-ending cycle, and in the end defeats the cost-saving effects of virtualizing systems. 4
Restore Order Implementing a well-designed strategy to manage your virtualized systems is advised in an effort to bring order to the all too familiar chaos. For each item below, we've provided a working solution that, if implemented, can lead to resolving the problem and reducing the associated costs. - Do you know what divisions of your organization are using cloud resources today? Whether you call it 'shadow IT' or just getting the job done, developers and business segments can easily initiate a new virtual computing environment whenever they need one simply by using their credit card. Providers like Rackspace, AWS, DropBox, and many others make it easy and inexpensive to set up computing services in a matter of minutes. If you don't know what and how many of these are being used in your enterprise, you are missing out on understanding the full range of computing needs, and being able to make the best decisions and recommendations for the company. Even though those services may be initially inexpensive, the costs can escalate and may become less cost effective than your internal resources. VMware s vcac (vcloud Automation Center) for cloud computing can give you a single pane of glass to manage and deploy virtual resources whether they are on-premises, deployed through Amazon's AWS, in VMware's vchs (vcloud Hybrid Service), and in other vcd (vcloud Director)-enabled clouds. Being aware of what is being used by what divisions, and what the utilization and costs are, you can make recommendations that can consolidate resources and reduce costs. - Does IT own all cloud instances, or do departments seek out their own? 5
Make sure that cloud-based resources are tracked to their Line of Business (LOB) owner so that you can report on cloud hosting costs on a department-by-department basis. Computing resources are becoming less expensive, but there are still significant costs to acquiring, deploying, and managing the systems used to run the enterprise. Providing periodic or on-demand reporting of resources used by each department, along with the costs of each makes it easy to account for costs incurred by IT. For example, IT needs to be able to show these 20 VMs support Accounting, these 15 support Marketing, these 10 support the entire company as supporting documentation for their cloud service costs. - Can you tell what resources your cloud instances are actually consuming? Are they too big, too small, or just right? Without making detailed inspections of every computing resource in the virtualized environment, it's nearly impossible to determine whether each virtual instance is deployed at its maximum efficiency. Tools like vsom (VMware's vsphere with Operations Management) verify that appropriate resources are allocated to business critical workloads. - Is your ability to spin up a new cloud resource measured in minutes, days, or weeks? Creating a new computing instance should be fast and easy. Your departmental users know that they can select an online resource from several vendors and set up what they need in minutes. They expect to be able to get the same results from their internal IT department, and when they face delays in getting what they need, they may find it easier and faster to use external sources. Ensure your cloud deployment methodology uses streamlined workflows with as few decision making checkpoints as possible to deliver what is needed in the timeframes they are required. 6
- Do your cloud resources get built by hand, or by a repeatable and automated deployment solution? Configuring resources customized for each deployment doesn't make sense when it's possible to set up configurations that account for 90% of the needs and can be generated automatically based on an existing configuration definition. Creating new instances manually contributes to the delays that cause departments to seek external resources. Make sure the most high touch human interaction in your cloud deployments consists of having the decision maker click Approve in an email. Your Next Move As cloud based systems become the norm for enterprise IT, be sure to stay ahead of the adoption curve by understanding the current needs within your enterprise as well as what you've already deployed. Creating new virtual instances makes it easy to quickly satisfy the needs of your business units that is, until you lose track of them and resource utilization interferes with normal operations. The key to staying on top of new deployments while controlling existing resources is having an effective management plan in place that doesn't add yet another layer of complexity to your growing environment. Stop the potential chaos with a sound and sensible management plan. 7
About egroup Providing customers across the southeast with a distinct competitive advantage since 1999, egroup is a visionary technology solutions and services firm with the ability to execute swiftly in a rapidly changing information technology landscape. Our solutions drive customers cost containment, revenue growth, and service objectives by addressing challenges associated with mobility, access to critical applications and data, and security. egroup s expertise and core competence is focused in three distinct areas: Cloud Services (public and private), Application Services, and End-User Computing. To deliver these solutions, egroup has strategic partnerships with industry leaders that include VMware, Cisco, and EMC, along with Citrix, Microsoft, and Trend Micro, and a highly-skilled workforce trained in all of the latest developments in information technology. For more information, visit www.egroup-us.com 8