Including Internet Load Balancing And Redundancy In DR Strategies

With VoIP continuing to demonstrate robust growth and the SaaS market projected to expand by twenty-two percent this year, companies are relying more on IP-based technologies for their communications and for good reason. IP in the form of VoIP is cheaper than traditional telephony, with its more expensive T1 and T3 lines. Internet-driven connectivity as delivered by wireline, cable, satellite and wireless ISPs also offers great flexibility.

Unsurprisingly, many small and medium-sized businesses (SMBs) plant IP-based technology squarely in the middle of their disaster recovery plans, believing that there is always the option of  sending people home to work in the event of  a company communications outage--but what happens if the Internet itself  fails? "There is a growing trend for companies to consider this question--and to look at a host of options, often in the form of a combination of wireline and wireless ISPs," said Darin French, Vice President of Business Development at xRoads Networks, a developer of cloud optimization appliances.

One such company is Town and Country, a Palo Alto-based SMB that has been doing business for over twenty years and provides premier placement services for nannies and home service workers. The company's thirty-five employees screen candidates and place them with clients. Much of this work is conducted by phone. Because telephony and strong network communications are vital, Town and Country wanted to ensure that calls were never dropped and that call quality was consistently high. "We originally had a single Internet connection for all of our offices and remote workers," said Stanley Hutchinson, Town and Country's IT Director. "When the ISP experienced problems, we could encounter a high failure rate on our data connectivity with the data base we used for the core of our business. This was unacceptable, so we began to consider a multiple load balance and failover VPN with tunnels to our offices, and to reconsider how we were using ISPs."

Located in an earthquake zone, Town and Country remembered lessons learned from Hurricane Katrina. Companies in the New Orleans area during Katrina were able to maintain communications by using wireless ISPs at a time when wireline communications were severed. "We wanted to have multiple ISPs for purposes of redundancy," said Hutchinson. "In making the decision to use multiple ISPs, we had both failover availability and the ability to load-balance traffic whenever we needed to, by automatically routing to ISPs that had additional capacity." The "mix" of ISPs that Town and Country used for the critical "last mile" of corporate communications at its Palo Alto headquarters and its San Francisco field office included at least one wireless ISP at both locations, as well as wired ISPs.

Town and Country's telephony network features an IP PBX at its Palo Alto headquarters connected to a local area network. The IP PBX takes all incoming calls to Town and Country from the public switched telephone network (PSTN). Town and Country's load-balancing and failover configuration intelligently routes voice and data traffic via three wired and wireless ISPs in Palo Alto. A branch office in San Francisco has its own local area network and VPN/firewall and uses a second network switch to route traffic through routers to two other ISPs, one wired and one wireless. The third ingredient in Town and Country's network consists of remote teleworkers who have IP phones with built-in VPN tunnels that seamlessly connect them to the headquarters PBX, with the full complement of multiple ISP routings and redundancies that in-house corporate users have."As companies become more savvy with IP technology, they are aggressively adopting best practices that further their networking and communications options at economic price points," said xRoad's French. "To address disaster recovery and failover, they are employing secondary ISP connectivity through broadband and wireless. In the area of voice and data load balancing, they are not automatically upgrading from T1 to T3 lines, as they might have in the past, but instead are considering a variety of broadband, wireless, cable and DSL alternatives that are able to give them bandwidth at much lower prices. They then provision their high performance applications over their best-performing ISPs, and route email and general Web access over other broadband connections."

i2spc  Vice President of Engineering Mark Langlais seconds this strategy. "We are running a SaaS supply chain business with a worldwide clientele, so we have to be up and running 24/7," said Langlais. "This means that we use multiple data centers with redundant servers and both UPS and generator backup. Just as essential for us are redundancy, failover and load balancing capability in our IP-based communications."