• Home
• GOVERNMENT CXO
• E-GOVERNMENT
• CITIZEN ENGAGEMENT
• TAX AND REVENUE MANAGEMENT
• GOVERNMENT PROCUREMENT
• GREEN GOVERNMENT
• GOVERNMENT DATA MANAGEMENT
• GOVERNMENT SECURITY
• GOVERNMENT CLOUD
• GOVERNMENT ANALYTICS
• GOVERNMENT GIS
• CONNECTED GOVERNMENT
• DIGITAL INCLUSION
• EDUCATION IT
• HEALTHCARE IT

Government Data Management, Technology

Meeting SLAs in Virtual Data Center Environments

25 January 2010

Virtualisation is on every CIO’s radar, and for good reason. It is widely considered to be one of the best ways to reduce costs and improve operational efficiency in IT, and the economic downturn has increased its prominence. After all, which CIO isn’t under pressure to do more with less today? VMware, with its commanding market lead and more than 120,000 customers is the virtualisation solution most often turned to.

While the server consolidation benefits of VMware are evangelised relentlessly, there are at the same time a number of consultants and pundits warning of the dangers of rushing in to virtualisation . Of course they are right, like most other things in IT, virtualisation needs planning and attention to detail to ensure success.

Rather than look at virtualisation as something potentially beneficial but with its own share of traps and pitfalls, it is more helpful to look at it from the perspective of a Service Level Agreement (SLA). Why focus on the SLA? Simply put, the SLA formalises the promise that the CIO makes to the business that he or she will provide the applications and information processing systems the business needs to run effectively and to be competitive in the marketplace. As such, the SLA provides an excellent map to the areas requiring special attention when virtualisation is deployed.

One of the most important elements of SLAs is the requirement for these business applications to be highly available, often 24x7. In order to guarantee this level of availability VMware provides several innovative High Availability (HA) capabilities, but these must be combined with corresponding HA facilities in storage and networking components of the infrastructure to deliver on the availability requirement. In a sense VMware-driven server consolidation causes all eggs to be placed in one basket, so at the physical infrastructure layer IT needs to ensure they are providing a very reliable basket. For consolidated VMware environments the use of storage arrays with five nines reliability (99.999%) is no longer a luxury, it becomes a basic requirement.

Similarly, the SLA addresses the expectation that the important applications will execute with sufficient performance, regardless of workloads caused by other applications. Meeting this requirement again requires coordinated technologies at the virtual and physical infrastructure levels. VMware provides sophisticated methods to ensure that applications get the resources they need, and that bad behavior from one application does not affect the others. To truly guarantee this lack of interference the solution needs to be extended to the centralised storage layer, ensuring that IO contention for storage resources can not cause application performance to deteriorate.

It is also necessary for these applications and their data to be recoverable in the event of equipment failures or user mistakes. Because VMware virtual machines are essentially indistinguishable from physical machines the current backup and recovery applications will work without change. However, the high IO loads of backup operations can become a bottleneck that will limit the degree of consolidation that can be achieved. Many organisations find that by employing newer backup technologies this bottleneck can be eliminated and significantly higher consolidation ratios can be achieved. The business case is easy to make and compelling. Solutions employing deduplication technologies and special provisions for backing up entire virtual machines at once can significantly raise backup and recovery service levels.

Another element of the SLA revolves around the handling of disasters. Should a disaster occur, IT must contribute to the continuity of the overall business by making sure these applications and data can be made available from an alternate site with minimal downtime. VMware’s Site Recovery Manager (SRM) coordinates and automates much of this process, and by eliminating the need for symmetrical hardware configurations between primary and recovery sites can significantly lower the cost of disaster recovery. In addition it can raise service levels by extending protection to applications for which previously the protection costs couldn’t be justified. Importantly, SRM leverages the remote replication capabilities of storage arrays to effect disaster recovery. Once again the quality and performance of the capabilities offered by VMware in the virtual domain are highly dependent on the characteristics of the physical infrastructure and storage systems in particular.

Increasingly compliance and governance requirements place demands on security and information retention practices that must be addressed. Special-purpose storage systems dedicated to long-term retention must be compatible with applications in virtual environments, and in some highly regulated industries the entire operating environment, both virtual and physical must be certified. Long-term archiving of virtual machines along with their data sets can overcome information retrieval problems when, years or sometimes decades later, applications and operating systems that are then in use can no longer read historical information.

For IT departments systems monitoring and management, diagnostics, capacity planning and chargeback are all essential elements of delivering high service levels to the business, and all are impacted by virtualisation . As a result there is a need to adopt end-to-end management tools that can address both the virtual entities and physical resources in the datacenter and understand how they relate and interact in delivering a service.

Our experience with customers has shown that starting with an SLA perspective can provide a manageable beginning to virtualisation projects, and can also serve as a check for virtualisation projects already underway. It can be a valuable guide to making crucial decisions regarding the physical infrastructure that hosts your organisation’s virtualisation platform. Have you evaluated your virtualisation initiatives through an SLA-centric lens? You might be pleasantly surprised how this approach can yield a roadmap that ensures a smooth deployment of virtualisation and maximal achievement of the financial, operational and business benefits that virtualisation offers.

---