Chilewich Sultan Weaves Together a Custom Disaster Recovery Solution
Chilewich Sultan benefits from a third-party hosted and maintained Power Systems server, necessary network support and a near-real-time backup environment.
Image by Robert Sentell IT Manager, Chilewich Sultan - Photo by Ben Rollins
By Jim Utsler04/03/2017
|Customer: Chilewich Sultan LLC
Headquarters: New York, N.Y.
Business: Designer, manufacturer and distributor of high-end woven products
Challenge: Grappling with a precarious disaster recovery solution
Solution: Partnered with several IT support companies to design a largely effortless backup and recovery infrastructure
Hardware: An IBM Power Systems server
Software: Maxava’s disaster recovery as a service cloud high availability solution and Avalara AvaTax
Design is everything at Chilewich Sultan LLC, creator and manufacturer of a furniture collection and high-end textiles for table and floor mats. This attention to detail also applies to its IT infrastructure, especially in terms of disaster recovery (DR).
That wasn’t always the case, however. Until last year, backups occurred daily and were sent to a third-party data center with a recovery time objective of up to 24 hours. A disaster could result in “48 hours of work down the drain,” says Robert Sentell, IT manager, Chilewich Sultan.
Now, thanks to its use of Maxava’s DR as a service (DRaaS) cloud high availability (HA) solution, Chilewich Sultan no longer worries about this worst-case scenario. Although Maxava typically offers this technology in conjunction with business partners that supply the hardware infrastructure, it made an exception in the case of Chilewich Sultan.
No matter how the solution is implemented, it can be a blessing to organizations that don’t want to fret over the possibility of data loss. Chilewich Sultan, for example, can now recover from a disaster within four hours, and all transactions posted before the disaster will have automatically transmitted to a partition on a Maxava-hosted IBM Power Systems* server.
Hit by a Hurricane
In 2000, Creative Director Sandy Chilewich founded the company, which first supplied woven extruded yarn products for the home, including placemats and tote bags. In addition to manufacturing and distributing those items, the New York, N.Y.-based company also introduced cubes—a line of functional furniture.
Based on its early sales success through retail stores, the company examined commercial applications for its products with Sandy’s husband, CEO Joe Sultan, who was trained as an architect. They spearheaded Chilewich Contract, which specializes in selling woven textiles for wall-to-wall and tile flooring to architects and designers.
This was followed by Chilewich Hospitality, which produces woven textiles for hotel and restaurant dining spaces; Chilewich Wholesale, which works with specialty and department stores to distribute consumer products; a licensing agreement with The Shade Store to produce a collection of roller, vertical and panel-system shades; and its first retail space, which opened in 2014 in New York City.
Most Chilewich Sultan products are produced in and distributed from Chatsworth, Georgia, but it also has an international distribution hub in Spain. Some web design takes place in the company’s headquarters, but its core IT department—which comprises three employees who oversee the operations of its proprietary ERP application—works in Chatsworth.
The company’s original Windows* on Intel* computing environment was based in New York until Hurricane Sandy hit in 2012. Widespread power outages left Chilewich Sultan as an unfortunate victim. Although the server itself was OK, it couldn’t be booted.
“We might be down for four hours—or less—but we wouldn’t lose our final transactions. And once we’re back online on the mirrored partition, we can immediately get back to work.”—Robert Sentell, IT manager, Chilewich Sultan
Sentell recalls how the company’s founder and CEO took matters into their own hands: “Sandy and Joe went to the office and climbed eight flights of stairs in the dark to retrieve the server, walk it back to their apartment and plug it in. It powered on, but it still wasn’t usable because nobody could log in. They were essentially down for five days until the power in their office was restored. Things might have been different if a redundant system was in place at another site.”
Following Hurricane Sandy, Chilewich Sultan considered its options, which included moving from x86 to another platform, largely because it was outgrowing its existing system. With the guidance of its COO, the company chose a Power Systems server that it subsequently moved to Chatsworth.
Chilewich Sultan also invested in a more robust DR solution. Before 2016, the company ran a daily virtual DR backup at 1 a.m., which was then sent via the cloud to a hosted DR site located out of state. This arrangement caused several challenges, not the least of which were the potential loss of data and delayed recovery times.
“One of the big problems was that we were only backing up once a day. Another was that the service-level agreement (SLA) with the provider was a 24-hour recovery time. As a result, we could not only lose a day’s worth of transactions but also potentially be down for another day waiting for the vendor to recover on its system,” Sentell notes.
Another challenge involved communications. After the company lost a T1 line during a heavy rainstorm, it had to revert to a redundant telephone line. Because the quality of that line was so poor and data transfer times were so long, a few nights of backups didn’t complete, which left the company facing days of makeup work and lost sales if its system had crashed.
Chilewich Sultan could have stabilized its DR environment by purchasing another Power Systems server and mirroring its production system to the backup system, but the overhead involved in terms of both hardware and software cost, and IT administration, would have been cost-prohibitive. It also would have compounded another issue: “Our computer room isn’t secure,” Sentell explains. “No lock is on the door, it’s located about 30 feet from a highway and the server was on an external wall. If a truck were to drive off the road and hit it, you can imagine how that might turn out. That was another key factor in our decision to redesign our DR infrastructure.”
To address these challenges—once-daily backups, a 24-hour restoration SLA, poor communication lines and a precariously located Power Systems server—Chilewich Sultan worked with several partners to better protect its core computing environment.
Maxava suggested moving to a more streamlined DR solution that uses the cloud to push production data to a partition on a Maxava-hosted Power Systems server. The goal was to leverage Maxava’s HA and DR expertise while, more importantly, moving away from daily backups to nearly real-time backups and improving rollover times. In initial testing, rollover time dropped to merely hours, based on a four-hour SLA, a marked improvement over 24 hours.
Meanwhile, the company moved its Power Systems server to a Peak 10 hardened server-hosting facility. This offered several advantages, including reliably redundant communications lines, a more secure server environment and on-site support, which was key, given Chilewich Sultan’s lean IT staff.
In-Touch Computer Services Inc., which oversees the company’s networking and PC maintenance as well as other IT resources, worked closely with both Maxava and Peak 10 to ensure that all of the expected IP addresses were configured properly.
With these three partners working in concert, Chilewich Sultan soon had a bulletproof DR infrastructure in place. Transactions on the company’s production Power Systems server are almost immediately recorded to the identically configured Maxava-hosted partition. If any synchronization issues arise, Maxava contacts Sentell, who reaches out to In-Touch, which in turn huddles with Peak 10 to configure the network to point Chilewich Sultan to the DR partition.
“This process, aside from the initial phone calls, would largely be invisible to us,” Sentell notes. “We might be down for four hours—or less—but we wouldn’t lose our final transactions. And once we’re back online on the mirrored partition, we can immediately get back to work.”
Although Chilewich Sultan hasn’t yet experienced an actual outage, it did implement an application—the sales tax calculator Avalara AvaTax—and attach it to its DR backup. Maxava only spent approximately one hour to add the associated library to this backup.
Focused on Work
Rather than continuing to grapple with a tenuous DR infrastructure—and tempting the fates—Chilewich Sultan has a largely hands-off solution that includes a third-party hosted and maintained Power Systems server, necessary network support and a near-real-time backup environment.
“We couldn’t have picked better partners. Now we can focus on what we do best, which is designing, manufacturing and shipping our products,” Sentell says.
Jim Utsler, freelance writer and former senior writer, has been writing about technology since the mid-1990s.
See more by Jim Utsler