Can the Cloud Take on Disaster Recovery? One City Says Yes
July 10, 2014
As someone who talks a lot about redundancy, I was troubled by the thought of talking about the cloud with regard to disaster recovery. City of Asheville (NC) CIO Jonathan Feldman, though, was impressed by the ideas and concepts put forth by CloudVelox. Like every good CIO, however, he started small with the company, migrating important but non-essential elements of the systems under his control, before moving forward with growing capacity.
I recently interviewed Feldman, asking him which applications Asheville considered to be essential. He pointed to their ERP system. He told me that the city has a strong IP infrastructure, and that in migrating their redundancy their ISP was concerned about the surge of data beyond the city’s traditional network, and thus immediately notified the city of the change of direction. Since then, though, the redundancy has been incremental and efficient.
Next I asked, “How does this tie in, if at all, with Asheville’s Emergency Incident Control Center plans”, to which Feldman pointed out that the COOP (Continuation of Operations Plan) identifies critical systems that must be set up for disaster recovery. This new method of recovery allows Asheville to extend protection to systems that don't qualify for COOP but are still important to stakeholders that don't have funding for traditional DR.
I followed up, enquiring as to whether Asheville had looked at alternatives, and Feldman assured me they had, but emphasized how impressed he was with CloudVelox’s rethinking of how DR could be performed. He also said that as it is a new system it has little impact on the existing network.
Finally, I then asked him about the impact of the new system, and also enquired as to how much management and maintenance are involved. He told me that the the system has been operationalized to test one system per quarter, with an expectation of one to four hours per exercise.
I offer this quote from a press release which outlines the major benefits of CloudVelox as they associate with the Asheville sports / event arena:
“The City of Asheville's IT Services department manages the nerve center for critical city services, from water to sanitation and key facilities, including the 7,500 seat US Cellular Center. Like many cities, disaster recovery excellence has been a challenge for the City of Asheville; protection was limited to critical apps, and their disaster recovery site was at an undesirable site two blocks away from their data center.
“Recently, the city faced challenges finding traditional disaster recovery funding for support for two key apps, one for asset management and the other for arena point of sale transactions. The asset management app is used by the city to track work orders and maintenance and to manage physical assets. An outage required ‘restore from backup’ with possible loss of one day of data and increased difficulty tracking work orders. The point of sale app supports event ticket sales at US Cellular Center. An outage during an event would impact sales volume until the system could be restored from backup.
“A big barrier to disaster recovery protection for these apps included substantial capital outlay requirements. By automating critical processes, CloudVelox was able to help the Asheville team extend disaster recovery protection to those important apps and enjoy a superior geographic operating model while also improving key disaster recovery objectives.
"‘The CloudVelox platform gave us the game-changing ability to deploy our important apps on AWS for disaster recovery with minimal budget impact compared to traditional disaster recovery solutions, which either didn't support our physical systems or required prohibitive fixed costs for services,’ said Jonathan Feldman, CIO of City of Asheville, North Carolina. ‘What we accomplished together should serve as an example of the higher standards of service and system protection now possible in state and local governments.’"
I wholly expect cloud services will become more integrated into our future systems, and with this in mind I expect DR will become a growing portion of Fog Computing.