"Big data" storage continues to hit wide area networks (WANs) hard as enterprises attempt to replicate large, unstructured...
data files across multiple locations. Large enterprises -- such as those in the entertainment and media industries -- that must replicate large graphic files, videos and data sets across multiple locations struggle with bandwidth constraints and WAN latency.
WAN latency concerns: Big data can't be ignored
To address these big data WAN latency issues, Silver Peak Systems has expanded its partnership with storage vendor EMC. EMC has tested and qualified Silver Peak's physical and virtual WAN optimization appliances with EMC's Isilon network-attached storage and SyncIQ data-replication products.
Silver Peak and EMC claim that together they can improve replication performance ninetyfold with SyncIQ and reduce the bandwidth requirements of this replication technology by 99%. The partnership -- which builds on the two companies' existing storage-replication relationship -- offers faster replication of big data storage across distances and bandwidth reduction for EMC Isilon customers without increasing costs, said Marc Trimuschat, vice president of business development and alliances for Silver Peak.
EMC, which acquired Isilon in 2010, has been focusing on making the technology meet the big data storage needs of large enterprises in entertainment, oil and gas, and several other niche vertical markets that have to deal with big data regularly. "Big data is becoming much more core to these enterprise applications," Trimuschat said.
Silver Peak is helping Isilon customers access this big data from remote locations, ensuring that it is replicated in a timely fashion, he said.
"[Enterprises] think that bandwidth pipe size is the only problem, but latency and data loss affects throughput," he said. "With the addition of Silver Peak, Isilon customers can now replicate data in six minutes, as opposed to a full day."
More on WAN latency:
How to measure your enterprise's WAN latency
Enterprise trends shape WAN performance
Meeting the need for WAN speed
Big data, different considerations
Enterprises must prepare the WAN to handle large-file access and backup across distances, as big data introduces new requirements for replication. Enterprises don't want to worry about "where the big data is," said Dave Bartoletti, senior analyst at Forrester Research. "[EMC's] Isilon is a powerful and scalable file-based solution, and [EMC customers] will be able to replicate the data very quickly, wherever it needs to be. Taking distance out of the equation is a good thing."
"If I need to work on a piece of data in New York and it was produced in California, I have to replicate that data across the country and that might take an hour. But when you can accelerate that process to happen in minutes or seconds across storage systems, it reduces the delay time for the information you need to be at your fingertips," he said.
Scalability will also be a major differentiator in the big-data market going forward -- a trait that Silver Peak has, noted Paula Musich, senior analyst of enterprise networking and security at Current Analysis.
The big-data phenomenon is still relatively new for many enterprises, Musich said. But the larger an enterprise is, the more likely it is to benefit from big-data replication offerings.
Speeding replication between storage arrays can boost reliability for the enterprise, whether large or small, Bartoletti said. Having data on-hand, on-demand, faster file sharing and reduced waiting time before a remotely located enterprise can start to process data is crucial for business continuity.
The key focus moving forward will be file transfers and big-data replication as enterprises continue to spread out, he noted.
"As enterprises look to consolidate data in a central storage location or even move storage to the cloud, I think that WAN optimization and the network becomes even more important across all enterprise locations," Bartoletti said.
Let us know what you think about the story; email: Gina Narcisi, News Writer