How Enterprise Data Storage Improves Database Availability

Most of my blogs are written with storage administrators in mind, with tips on how to keep data center storage running smoothly. However, in this blog, I am going to address the people that are the chief beneficiaries of quick and resilient storage – the administrators of enterprise application software (EAS).

 

EAS applications cover a wide range of workloads, and a never-ending supply of acronyms including ERP (Enterprise Resource Planning), CRM (Customer Relationship Management), POS (Point of Sale), BPM (Business Process Management), HR (Human Resources), AP (Accounts Payable), AR (Accounts Receivable), SOX (Sarbanes Oxley), as well as messaging, i.e., e-mail. If you are the one of the people responsible for keeping applications such as these running effortlessly within your organization, this blog is for you!

 

I’d venture to say that most EAS administrators are not very aware of the underlying storage that powers their applications. Sure, they want fast, reliable storage - and for that, they rely on their IT department to make the right decisions to provide this. However, I contend that application administrators should be part of the storage decision-making process, because, unlike servers and networking hardware – implementing the right storage infrastructure can actually fix broken applications and make them run right when everything around them seems to go wrong.

 

In the minds of EAS administrators, there are three big problems that cause things to go wrong:

  • Inconsistent application performance
  • Lack of application availability
  • Application data corruption

Let’s take a look at each and see how NetApp addresses them.

 

  1. Inconsistent Application Performance

Most enterprise applications, particularly those hosted on databases, experience periods of performance spikes. Spikes can dramatically slow down application response times, and usually occur during predictable events – a special promotion, a seasonal pattern, end-of-month and end-of quarter reporting periods, etc. Although these periods are predictable, they are not necessarily planned for. To prevent performance spikes from affecting operations, NetApp offers two solutions:

 

  • NetApp Virtual Storage Tiering is an automated method for moving hot data to a flash tier. As data is accessed, it is automatically promoted to flash. Hot data will remain in flash, while cold data is evicted. In this way, the effect of performance spikes can be minimized.
  • NetApp DataMotion is a method of seamlessly moving entire applications between storage tiers in a clustered storage environment. For instance, a retail company might require accelerated POS application performance during seasonal buying periods. To accomplish this, the entire application could be migrated (with DataMotion) from HDD storage to flash storage – while the application is in operation and without disruption. Once the seasonal period ends, the application could be moved back to traditional HDD storage. Even better, this entire “performance migration” operation could be triggered automatically using NetApp Workflow Automation.
  1. Lack of Application Availability

Enterprise storage vendors strive to design resilient storage systems with redundant components and self-healing capabilities, and NetApp is no exception, with its EF-Series and FAS storage arrays demonstrating 6 nines field reliability.

 

However, unplanned downtime due to component or software failure is only one aspect of continuous application uptime. Scheduled downtime for system maintenance should also be eliminated. NetApp accomplishes this through a well-thought-out approach to nondisruptive upgrades (NDU). NetApp NDU enables upgrading and maintenance of any aspect of the storage system by insuring that maintenance-related I/O interruptions are brief enough so that applications continue to operate without the need for downtime, maintenance, or user notification. The end result - 24x7x365 continuous availability of critical enterprise applications.

 

  1. Recovery from Application Data Corruption

According to a DBA I recently spoke to: “Database corruption doesn’t happen all that frequently.   But, when it does, it is really an unpleasant thing to deal with”. When a corruption event occurs, rather than being catastrophic, it is more likely to be a database problem that prevents access to a certain table within the database. The first step is usually to apply the database vendor’s consistency checker, which I am told will correct the problem about 60% of the time, but often with some data loss within database tables.

 

If the checker cannot fix the corruption, or the fix resulted in data loss, the next step for the DBA is to identify when the corruption occurred. DBA’s do this by rolling back to earlier snapshot copies until they find a clean (non-corrupt) copy. If they were fortunate enough to have stored the data on a NetApp system, they can spin off a clone of the earlier database version and apply redo logs to bring that DB up to date. Once they have this copy, they can compare the preserved database with the production version and determine which rows and columns are missing, and copy the missing data to the production database. The final step is to remove the clone copy and breathe a sigh of relief.

 

Summary

EAS administrators are extremely risk averse. Who can blame them, when the entire organization rests on their shoulders. Of course they’d all like their applications to run faster with endless endurance, but they are not willing to take one iota of risk in order get there. NetApp’s portfolio leverages proven platform architecture with tools designed specifically to mitigate risk while accelerating application speed and availability.

 

Resources:

Best Practices for Oracle Databases on NetApp Storage

Best Practice Guide for Microsoft SQL Server with NetApp EF-Series

NetApp Extreme Performance Solution for Oracle Database