In 2000, when international freight forwarder Röhlig and local logistics manager Grindrod merged, Röhlig-Grindrod inherited the joint IT landscape. The new company improved that landscape in several ways.
In 2005, the company switched from hosting its own servers to a hosting provider. It was migrating the mess of various Windows desktop systems to virtual desktops and moving onto a MPLS-managed wide area network, to effectively make `local data calls’ from its branches to its hosting provider's local sites. This meant Röhlig-Grindrod no longer had to rent costly dedicated land lines from head office to its eight national branches.
But the inheritance was still difficult to keep going smoothly.
Data management was an absolute nightmare, says Abdullah Kahn, Röhlig IT manager, because viruses in the archaic e-mail system used at that time created a huge risk for the business. Viruses had caused havoc several times already.
The application users at the branches had erratic experiences, says Kahn. At times they would have fast application delivery, sometimes significantly degraded.
“All branches were a pain,” says Kahn, “but the big ones, Johannesburg and Durban, were a bit more of a nightmare to manage.”
At the time it wasn’t possible, from head office, to figure out where the branches' bandwidth bottlenecks were.
“Let's say I am printing a huge report in East London. East London would have a little Diginet line, nothing significant, because there are few users working there. When I print that report, the other users are now experiencing a slower performance on their applications, because of the print job,” says Kahn. “We could not identify printing traffic, from e-mail traffic to business-critical Sequel Server application traffic.”
Yet another malicious virus outbreak, via e-mail, in 2006 forced the turning point.
Find the gridlock
Röhlig-Grindrod asked Mike Zietsman, Opteq South Africa CEO, to investigate the management of its wide area network (WAN). Zietsman found problems `that any company faces’.
Röhlig had different types of devices to manage its security, its users, its WAN traffic, says Zietsman, as well as its network monitoring and IT support.
The devices were installed by major industry players, but Röhlig managers were unable to get consolidated reports to quickly make sense of what was going on. Each device had its own report.
“This company supporting this service blamed that product from another company because it wouldn't integrate,” says Zietsman. “A mess we normally deal with.”
Opteq plugged one of its devices, with bandwidth management module activated, into the Röhlig WAN.
“What the Opteq bandwidth manager found,” says Kahn, “were a number of virus sources.”
Röhlig could tackle the virus problem then, even without IT support at each branch.
“With this device, we could identify the IP address of the virus originator, and deal with it there and then,” says Kahn. “With the branches that had an IT guy, we could ask him to do the scan and the virus cleanup, right away.”
Then Opteq targeted the data traffic between head office and branches with the bandwidth manager.
“We shaped the business-critical systems, and gave them priority,” says Zietsman.
At this point Röhlig could limit the amount of printing traffic at a branch, says Khan, based on how much bandwidth the Opteq bandwidth manager allowed that branch. A big printout no longer prevented users of business-critical applications from getting their jobs done.
“When a user sits in East London, connected to the WAN, they don't wrestle for the same piece of wire that the e-mail and internet is wrestling for. It's like we put them in their own, dedicated traffic lane and nothing can touch that,” says Zietsman. “The performance is almost like sitting on the local area network (LAN), working off the server.”