When you think of the kind of applications that absolutely, positively, must not stop, you probably key immediately on those rarified high-performance systems fielded by financial industry giants. Think First Data Corporation, with its millions of credit card transactions that must be processed every few milliseconds. Think Bank of New York, whose deposits must be made to the Federal Reserve Bank by the appointed hour each day, or the value of the US currency might become unstable.
You know what I mean: the stuff you read about in brochures from EMC, IBM, HDS or the other mavens of "big iron solutions to big iron problems."
In all likelihood, the last thing that comes to mind is agriculture. If you aren't in the food industry, you might assume that the most technology needed by folks on the farm is, maybe, an e-book version of The Farmer's Almanac. The only big iron they require is the kind that bears the logo of John Deere or International Harvester.
The idea that short-term outage sensitivity could be associated with systems that support activities long measured by lunar phases and growing seasons seems pretty ridiculous. That is, until you talk to Daniel Hart, CEO of Serveon, in Wilmington, Del.
Hart will tell you some pretty frightening tales about the sensitivities of agricultural workflows and their increasingly mission-
We may not be talking currency destabilization here, but we are talking about the safety of the food supply. Serveon, which is part application software developer and part application service provider, offers large and small agriculture companies with what Hart claims is "the world's first food supply traceability platform -- from field to fork."
His clients depend on him to maintain real-time data warehouses that provide detailed logistics on the harvesting, transportation and spoilage of their product. And that doesn't begin to cover the problem of pesticide management, an important dimension of farming today. Partly in response to EPA regulations, partly in recognition of the lethality of some chemical compounds to life forms other than Caribbean Fruit Flies or other insects, pesticide tracking is another key focus for Serveon. (After listening to Hart describe the myriad challenges associated with pesticide records management, this writer will never, ever prepare vegetables or fruit for his family meals without scrubbing them first!)
Bottom line: Agriculture has some serious information requirements, and serious vulnerabilities to interruption in information access. So, even a small service provider like Serveon has a need for a "world class" information protection strategy.
The problem is, of course, that Serveon, like many smaller firms, has neither the budget nor the staff to support a big iron-based mirroring scheme. Instead, Hart has teamed with Zerowait, a high-availability network storage engineering company in Newark, Del., to come up with a common sense solution.
Design is key
Serveon's solution to high availability is enabled in part by application design. In many cases, database systems require a stateful client-server relationship that is prone to erred(erroneous?) or broken transactions if communications or server equipment problems arise.
In recognition of this fact, Serveon has made "stateless operation" a design criterion in many of its applications. Clients use a handheld computer to scan barcodes on produce crates, then insert the device into a networked cradle that transmits the stored data to Serveon's site in a batch mode. This "all-or-nothing" transaction, according to Hart, ensures that partial updates are not made. Any errors that occur in transmission simply abort the update process and trigger a resend of the data. Hart says that Virtual Private Networks are used to facilitate secure transmission.
Serveon hosts the server portion of its application on HP/Compaq Proliant servers with direct attached storage, "mostly because the gear has proven very reliable and you can get replacement parts readily." The company uses network-attached storage to host its e-mail systems, taking advantage of the e-mail recovery capabilities of his platform. Uninterruptible power supplies and a diesel generator provide protection against power outages, and a combination of Cisco Systems switch gear and NetScreen firewalling provide protection against network outages and security threats.
However, last year's power outage in Wilmington tested the mettle of the company's disaster prevention-focused strategy. Hart says that it underscored a vulnerability of his hosting business to interruptions in services resulting from external sources beyond his control. So, he turned to his engineering services provider, Zerowait, for an additional measure of protection.
Zerowait President Mike Linett reports that the proximity of the company to his offices (less than 10 miles) made application layer (layer-7) failover possible. He worked with Hart to deploy identical infrastructure to Serveon's in his own data center and established a load-balancing scheme between the two sites.
"We monitor the heartbeat of the Serveon systems," Linett says, "and failover to our equipment if anything goes wrong." He says that the process is so swift that only a secure socket layer (SSL) connection might be interrupted if failover was required, representing only a minor interruption.
Updates that are transmitted to Serveon's system are also transmitted to Zerowait, where they are applied to keep databases up to date. Secondary MX files from Serveon's e-mail system are transmitted to Zerowait on a routine basis and applied to the hosted e-mail, keeping the two in sync.
That the service provider and the engineering company are served by different power grids, different telecommunications central offices, and different network points of presence -- despite their geographic proximity -- helps to add an additional layer of protection to Serveon's operations. It also helps to keep the pricing of the arrangement on the ground floor for the service provider.
Bottom line: Sometimes the best data protection solution for the SMB is right around the corner. If your reseller/integrator has the competence, such arrangements could ensure the uptime of your business without the big price tag.
For more information:
About the author: Jon William Toigo has authored hundreds of articles on storage and technology along with his monthly SearchStorage.com "Toigo's Take on Storage" expert column and backup/recovery feature. He is also a frequent site contributor on the subjects of storage management, disaster recovery and enterprise storage. Toigo has authored a number of storage books, including "Disaster recovery planning: Preparing for the unthinkable, 3/e."
This was first published in February 2005