Dell Speeds Recovery Time by 99 Percent, Boosts Service and Efficiency, and Cuts Costs

“When peoplevisit our website, we want to sell them the best system for the most competitive price and, in the process, make sure that they receive excellent service. SQL Server 2012 helps us do that.”

Reinaldo Kibel, Senior Architect,Database Management Practice,

Infrastructure Technology Delivery Services, Dell

Dell sought a solution that could enhance the availability and performance of mission-critical systems that support its website. In 2012, the company began to upgrade 27 clustered databases to Microsoft SQL Server 2012 Enterprise. By doing so, Dell has sped up recovery time from 3 hours to 15 seconds, achieved 99.99 percent uptime, and boosted performance. As a result, Dell can maintain excellent service,increase efficiency, and cut costs.

Business Needs

Dell operates one of the busiest websites in the world. With only a few physical stores, the company facilitates most sales and support processes online. As a result, its site supports about 10,000 concurrent users during peak times.

One of the key initiatives at Dell is to providea positive experience for online shoppers, which requires maintaining high availability and fast system responses. “If a customer spends time on our website building a system, and the site suddenly goes down, I don’t think that individualwould want to come back in three hours and redo everything,” explains Reinaldo Kibel, Senior Architect with the Database Management Practice, Infrastructure Technology Delivery Services at Dell.

To help achieve high availability on the 27 database servers that support the website, IT personnel mirror and cluster them in multiple datacenters in the United States. Before 2012, all of these systems ran on Microsoft SQL Server 2008 R2 Enterprise data management software and the Windows Server 2008 R2 operating system. Collectively, the databases include approximately 40 terabytes of data. If an outage occurred, IT personnel had to update network settings to connect clients to a secondary database—a process that took as long as three hours. Engineers worked with a Microsoft Services consultant to develop a custom solution to automate failover.However, it supported only one mission-critical system.

Dell wanted a simple, out-of-the boxsolution to automate recovery. It also sought an easy solution to provide real-time reports. Before 2012, IT personnel used database snapshots to support reporting. The snapshots were updated on a regular intervalevery few hours. However, snapshots did not contain real-time information and their maintenance created extra work for IT personnelin case of an outage. In addition, Dellwanted to streamline backup processes so that they would not slow down response times.Kibel says, “Because we manage daily backups on 5,000 different servers, occasionally a backup runs duringpeak times, and that affects the load on database servers, causing website performanceto be slow.”

Solution

To address these challenges, Dell began testing SQL Server 2012 in August 2010. IT personnel becameespecially interested in SQL Server 2012 AlwaysOn feature, which automates system failover and maintains active secondaries that can offload processes from production systems.“There is so much at stake when it comes to our website that people areextremely hesitant to make any changes to the underlying infrastructure,” says Kibel. “However, when we demonstrated a failover situation where we used SQL Server AlwaysOn availability groups, people were extremely impressed. AlwaysOn merges best-of-breed database mirroring and clustering so that systems automatically failover without requiring application changes—the database system gains control over the process.”

In September 2011, Dell engineers started to upgrade some critical systems to SQL Server 2012 Enterprise, including servers that support online shopping carts, configuration data for products being purchased, and software downloads for hardware drivers. Every primary database belongs to one availability group with one active secondary. IT personnel configure each availability group to use synchronous-commit mode so that database changes are made to the primary and secondary serversto prevent data loss. Engineersare also moving processes such as backups and reporting to the replicas.

By February 2012, Dell had upgradedfour core systems (including 12 servers) to the new solution. Dell expects to deploy SQL Server 2012 on all 27 systemsthat support website transactions by November 2012.

Benefits

With SQL Server 2012, Dell achieves high availability and speeds disaster-recovery time by 99 percent. It also enhances service, increases efficiency, and cuts costs.

Speeds Disaster Recovery by 99 Percent and Achieves99.99 Percent Availability

By upgrading databases to SQL Server 2012, Dell achieves 99.99 percent availability. If an outage occurs, customers might notice slightly slower response times, but they will retain their online shopping cart including configuration information. Similarly, people who are downloading software can complete the process even if a database outage occurs. “With SQL Server 2012 AlwaysOn, systems automatically failover in15 seconds, which is 99 percent faster than before,” explains Kibel. “Our vice presidents and top management are astonished by the faster failover times.”

Maintains High Customer Satisfaction

High availability on Dell.com is one of the key requirements for customer satisfaction. Fast system responses are another. By upgrading systems to SQL Server 2012, Dell can deliver more consistent website performance. “When people visit our website, we want to sell them the best system for the most competitive price and, in the process, make sure they receive excellent service,” Kibel says. “SQL Server 2012 helps us do that.” For example, he explains, “We can now use replicas to support backups of our mission-critical systems. By doing so, we offload tasks that typically consume 20–30 percent of systems’ CPUs, which improves system performance accordingly.”

BoostsEfficiency and Minimizes Costs

With its new solution, Dell can increase productivity, which cuts costs. “We can use SQL Server 2012 to help IT personnel so that they do not have to manage manual failover processes,” explains Kibel. “Being able to access data from a replica is another huge improvement that we realize with SQL Server 2012. And that’s because not only can I now avoid creating snapshots for reporting, but I can also generate real-time data in our reports, so that employees can work more effectively.”

Because IT employees have fewer processes to manage and the company can better utilize its infrastructure without sacrificing availability or performance, Dell increases its ability to grow and to innovate. “SQL Server 2012 is much more affordable and easier to manage than other leading database technologies,” Kibel says.“Our high-availability solution on the Microsoft platform is cutting-edge, and the technology is second-to-none.”