Apache corporation energizes exploration with NetApp storage and disk-to-disk backup

The challenge
Implement a storage architecture to enhance technical application performance, ensure recoverability and reduce costs

Managing exploration computing for a high-growth oil and gas company presents nonstop challenges. On any given day, Bradley Lauritsen could be setting up new seismic data-processing systems in Cairo, checking on storage requirements in Buenos Aires, integrating data from a recent Gulf of Mexico acquisition, or preparing for a hurricane hit in Houston. As team lead, Lauritsen knows that supporting the company’s global Exploration Computing department requires a storage architecture that’s scalable, flexible, and reliable. “Our job is to make sure that Apache geologists and geophysicists have the data they need when and where they need it. Any delay in access can directly impact the company’s time to discovery and ultimately its ability to get wells drilled.

“A few years ago, we recognized that our direct-attached storage infrastructure no longer delivered adequate performance or reliability for demanding seismic processing applications. It was particularly difficult to recover from disk failures – and there were many. One hardware failure resulted in the loss of 1TB of data. Although we eventually retrieved 98% of it from tape, the recovery process took two days, and during that time our users could not work with any of the data. That event really emphasized the need for an updated storage architecture.”

The solution
Standardize on NetApp solution for exploration computing

Goals of the Apache storage infrastructure project included improving performance for technical users and applications, ensuring data protection and rapid recoverability, and achieving administrative cost savings. The storage solution also had to accommodate high data growth and support future technical requirements, including a mixed environment of UNIX® and Windows® applications and cluster-based computing.

After an evaluation of competitive systems, the Apache team selected a NetApp solution. “NetApp delivered on all the essential requirements and offered unique advantages in manageability, including having a common operating system across the entire product line. The NetApp solution was easy to install, easy to administer, and required no special training.”

Lauritsen points out that the availability of a near-line storage solution was another advantage that NetApp offered. “We wanted to explore SATA-based storage platforms for disk-to-disk backup, as well as for applications like our Landmark ProMAX software that utilizes a lot of read-only files.”

Today, NetApp FAS series and NearStore® systems provide high-performance, high-availability storage for Apache exploration computing environments around the globe, including U.S. sites in Houston and Tulsa, along with international sites in Aberdeen, Buenos Aires, and Cairo. NetApp solutions serve as primary storage for a wide array of mission-critical technical applications such as Landmark GeoGraphix and OpenWorks and Schlumberger Petrel subsurface interpretation and simulation tools based on Windows.

NetApp NearStore solutions also provide file storage for the ProMAX application, online storage for local and remote-office backups, and off-site data storage for disaster recovery. NetApp SnapVault® software enables automated backup of both local and remote storage to NetApp NearStore systems.

At the Houston site, Apache leverages NetApp Snapshot™ technology to make near-instantaneous data copies every two hours. NetApp SnapVault software replicates Snapshot backups (from Houston and from remote offices) to NetApp NearStore systems installed at a CyrusOne managed hosting facility. After an initial full backup, the NetApp SnapVault software ensures both network and storage efficiency by transferring or storing only changed blocks. The CyrusOne installation serves as a warm site from which data can be quickly recovered in the event of a primary-site disaster. Apache maintains six months’ worth of backup copies at the secondary site.

Business benefits
Automated online backup, disaster recovery, cost savings and flexibility

“We originally leveraged the NetApp SnapVault software for fast, online backup of seismic data,” comments Lauritsen. “SnapVault streamlines backup processes and allows our users to very quickly retrieve data without having to go back to tape. Users also have the flexibility to roll back to earlier Snapshot copies. That’s very useful, for example, in horizon or fault interpretation processes – if one view doesn’t provide sufficient insight, a geologist can easily revert back to the original data and take a different tack.

“Prior to implementing SnapVault, if we had a major data loss, it might have taken days to recover a terabyte of data. Now we can recover it in just a few hours and with much better granularity – we can actually view and back up data at the file level. When we saw how well the SnapVault software worked, we decided to use it to back up all of our exploration applications data.”

Off-site disk-to-disk vaulting: economical disaster recovery
After initially replicating to a NearStore system within the Houston data centre, Lauritsen says the next step was to set up a remote site that would enable recovery of data in the event of a disaster event at the main Houston facility. Apache decided to move the NearStore system used for backups to the remote facility and replicate directly to it. “The whole process of setting up replication to the CyrusOne facility took only about four hours,” asserts Lauritsen. “When it was finished, we effectively had a warm site that would enable us to rapidly bring key functions back online – and continue operating the business – after a significant failure event.

“We’re one of the first independent oil and gas companies to implement a disaster recovery site – in the past, many thought that only the major oil companies could afford such protection. But the economies of the NetApp technology combined with the cost savings of backing up directly to the remote data centre facility have made it feasible to implement a second-site solution. Whether it’s another hurricane or simply a power outage at a remote site that causes a data loss, we can instantly access information at the backup site and restore it to any facility around the world.”

Fast, automated backup: protected data, administrative savings
“Before we implemented the NetApp solution, we were not able to complete all of our Houston backup processes – representing about 10TB of primary data – within our weekend backup window. And the reliability of backups was questionable. Because of the speed and efficiency of SnapVault software, we now perform full Snapshot backups daily and store six months’ worth of backups off-site. Backup windows aren’t even a consideration anymore, and we have all the time in the world to archive to tape, if needed.”

Lauritsen estimates the company has realized an annual tape-media savings of at least $50,000 to $60,000. “We’ve been working to eliminate tape as much as possible. When you’re utilizing 100 tapes a month at $100 each, it adds up. We’ve already saved significant administrative costs by not having to manage backups, pull tapes, move tapes around, etc.”

In terms of administrative resources, Lauritsen says Apache has realized savings at both central and remote offices. “At the remote sites, we no longer need an administrator to run backups – it’s all done automatically with SnapVault back to Houston. And again, because of the efficiency of SnapVault, we have been able to utilize existing T1 connections.

“But the bigger administrative savings have been in our department. We are a team of four supporting a lot of applications, servers, and storage for a global exploration computing environment. That’s why solution ease of use is key. We need a user-friendly solution that takes very little manpower to administer – we simply cannot afford a complicated environment that requires diagrams, constant training, and a lot of people to keep it running. Since we’ve deployed the NetApp solution, we’ve grown our disk capacity by some 700% and yet have actually reduced our department size by more than 30%.”

40% performance improvement: faster time to results
Lauritsen says that since deploying the NetApp solution, Apache has seen improvements in both reliability and performance. “After we implemented the NetApp systems, users actually complimented us on the performance gains they saw. We estimate that the NetApp solution immediately gave us a 40% improvement over the DAS environment. Better performance and better reliability have meant faster time to results as well as the ability to process more data.”

Scalability for growth and new computing models
The architecture of the NetApp platform ensures that the Apache exploration computing team will be able to keep up with the company’s rapid data growth as well as leverage new compute models to accelerate workflows and expand in-house capabilities. “Just last month we added 6TB of storage,” Lauritsen notes. “But not only do we need to be able to scale capacity, we need to be able to scale out performance to support expanding compute requirements. When we originally deployed the NetApp systems at the CyrusOne facility, we were very much focused on enhancing data recoverability while keeping costs under control. Today, we’re looking at the facility as a potential cluster-computing site. The expandability of the facility and of the NetApp architecture gives us a lot of options in implementing the technologies we need to stay competitive.”

© 2008 NetApp. All rights reserved. Specifications subject to change without notice. NetApp, the NetApp logo, NearStore, and SnapVault are registered trademarks and Snapshot is a trademark of NetApp in the U.S. and other countries. UNIX is a registered trademark of The Open Group. Windows is a registered trademark of Microsoft Corporation. All other brands or products are trademarks or registered trademarks of their respective holders and should be treated as such.