Critical Issues Facing the Data Center

Database administrators (DBAs) and the enterprises that rely on the complex systems they manage face a host of challenging issues, many of them only dimly understood. Many of these issues are posed by emerging or established business trends. While end-users are grappling with the implications of this evolving business and technology environment, the major database vendors and the integrated service providers (ISPs) that often implement large deployments are finding their traditional business and revenue models potentially undermined by these trends and other factors.

This article examines a sampling of these issues, discusses their impact on the business as a whole - and particularly on the data center - and suggests how to mitigate the associated challenges.

Issue 1: Open Source Software

Open source software is defined by the Open Source Initiative as, “A development method for software that harnesses the power of distributed peer review and transparency of process. The promise of open source is better quality, higher reliability, more flexibility, lower cost and an end to predatory vendor lock-in.”¹ This novel, collaborative approach is widely deployed across organizations. The trend has reached the data center, with solutions such as MySQL almost universally adopted in the start-up and online communities. Market drivers spurring the adoption of open source alternatives include the range of opportunities for end users in terms of licensing models and customized solutions that can be developed more quickly and deployed at a lower cost than traditional solutions. Because the open source community is growing and supportive, application developers with varied skill sets need not rely on traditional vendors to develop solutions to meet the market’s needs. This widespread adoption of open source solutions causes stress for traditional database vendors that are accustomed to collecting substantial up-front fees from customers.

But for DBAs and IT managers in start-up companies, open source software presents an obvious and significant advantage to its traditional counterparts because the cost of a MySQL deployment can be deferred until it generates revenue. This deferment of cost drives DBAs in the offline world to adopt open source solutions in order to reduce up-front costs and minimize total cost of deployment.

As open source software takes hold across the enterprise, adoption rates will rise in the data center as well, perhaps even incidentally. For example, a company may introduce an open source element like a blog on WordPress that is backed by MySQL, and suddenly it becomes an imperative part of their infrastructure.

Issue 2: Virtualization

Virtualization is often introduced to take advantage of economies of scale (thanks to more robust hardware and software) or to save energy. However, virtualization presents opportunities and challenges that impact the business as a whole, especially regarding storage, computing, networking and security.

Each of these affect the data center, but the most challenging impact when virtualization is adopted is the massive, usually one-time effort for DBAs who must port applications to the new configuration. When facing this significant surge in workload, the additional resources of an outsourced database administration service provider can help get the job done as a one-time task without having to add permanent resources.

Businesses seem to recognize the potential savings and efficiencies offered by virtualization, and purchasing numbers support this trend. The recent “Corporate Software Purchasing Trends” study by ChangeWave Research reveals that 18 percent of the close to 2,000 respondents to their survey said they “will increase purchases of virtualization software over the next 90 days,” which is up five points from the 13 percent recorded in January.²

From an industry perspective, virtualization undermines traditional licensing models based on a per-CPU pricing structure, which is an issue that all the database vendors face. Now that multiple installations can be run on the same CPU - something that dynamites traditional per-CPU approaches - vendors are struggling to create a new pricing model. One exception to this is Microsoft SQL Server, which is licensed on a per-socket basis. This allows companies to take advantage of multicore technology without breaking the bank.

Issue 3: The Greening of the Data Center

As individuals and companies focus on the need to adopt sustainable business practices and make environmentally conscious decisions, enterprise IT departments face not only allocating resources and purchasing equipment in concert with the company’s budget considerations, but also keeping the equipment’s near and long-term impact on the planet in mind.

Nowhere is this more prevalent than in the data center. Traditionally a significant energy hog, the data center has been a major cost center that did not efficiently make use of its resources and has generated significant energy issues and corresponding environmental concerns.

According to the McKinsey & Company report “Revolutionizing Data Center Efficiency - Key Analyses,” data center energy use doubled between 2000 and 2006, and by 2012 it is expected to double again.³ The study also noted that for many industries, data centers are one of the largest sources of greenhouse gas emissions. IT departments need to be cognizant of emissions standards and augment their data consumption accordingly.

The intensive resource requirements of the data center can largely be mitigated by virtualization, as previously discussed. However, internal and external incentives are needed in order for DBAs to implement such efficiencies. While DBAs can’t influence utility providers or government to create incentives, pressure can be applied internally to commit the organization to sustainable practices that benefit not only the data center but also the company’s environmentally conscious image in the marketplace.

Virtualization helps to quell the emissions concerns where alternate approaches to achieving high performance, like federated architecture, can make the environmental problem worse.

A federated architecture enables companies to take advantage of the ever-falling cost of hardware and software to run their applications across large numbers of servers. Applications are then architected so that a robust decision-making layer directs every task to the most convenient or available resource. Although this helps to improve efficiency and maximize performance, these additional servers generate additional heat, take up space and use more energy than with the virtualization approach.