It has been perceived wisdom that only large companies need or can afford high-availability solutions. Typically, these have been large financial services companies, parastatals, retailers, manufacturers and Internet service providers, to name a few.
A key factor in this thinking was the cost of high-availability solutions, which have typically been viewed as out of reach for smaller companies. Accordingly, these companies have got by with a modicum of downtime, and simply accepted it as "one of those things" and lived around it.
For a variety of reasons, that can no longer be the case. By any measure, some of South Africa's "smaller" companies are now substantial entities, and they cannot afford downtime. Most companies today compete on the basis of their business processes, and their computer systems drive and fulfil these processes.
So, if their systems experience outages for whatever reason, there are serious consequences:
* Manufacturing or distribution come to a standstill;
* As companies increasingly operate 24x7 and across national borders, they are seriously compromised when online shoppers are turned away;
* Microlenders are unable to fulfil and process the thousands of loans which are requested and granted each day; and
* Employees cannot access their e-mail - and studies show that inability to access e-mail causes more stress than a car accident.
By any measure, some of South Africa's "smaller" companies are now substantial entities, and they cannot afford downtime.
Raul Garbini is a director of Edgetec
Before discussing the solutions available, it is helpful to understand the two concepts of recovery time objectives (RTO) and recovery-point objectives (RPO). RTO deals with the length of time required to recover data, RPO with the completeness of data recovered. Imagine a graph with RTO as the vertical axis and RPO as the horizontal axis. At the very bottom left of the graph would be tape backup, and at the top right would be high availability, a process technically known as logical data replication plus switchover (LDR+Switch), which quickly moves users and processes to a fully mirrored secondary server for it to assume all or most of the functions of the production server.
Affordable
Currently, the perception of most smaller companies is that this high-availability approach is unaffordable. Fortunately, this has changed in recent years due to the following factors:
* Hardware: A second server with sufficient capacity will be needed to accommodate replicated data and production requirements. So this server may be required to run applications from the backup environment during downtime, planned or unplanned. Depending on business requirements, the business may need a server of similar capacity and performance. Whichever way, a second server is needed, ready to switch over at short notice. A second server today is not an insurmountable cost for most companies.
* Bandwidth: To have true disaster recovery in place, a company must have its second server located off-site, which implies enough communication bandwidth to address the volumes of data flowing between it and the production machine. This must include the I/O processing capacity of the backup server and the comms lines between the sites. As we have seen, the cost of telecoms has dropped significantly in recent years.
* High-availability software: This executes, manages and monitors the replication or mirroring of vital data to the backup server. It also moves users and processes to the backup server during downtime. Costs here include purchase cost and annual maintenance and installation and training. Again, the cost of such software has dropped in recent years.
* High-availability management: It is impossible to get around the fact that some level of management is needed on a daily basis. If the selected solution has autonomic features, it can reduce the amount of management, and attendant costs, quite dramatically.
In the next Industry Insight in this series, I'll unpack each of these savings areas in more detail.
* Raul Garbini is a director of Edgetec.
Share