Maximum Uptime vs Guaranteed Data Delivery
Posted on July 15th, 2008 by Eric MurphyI’m back from my short camping trip and am slowly whittling down the accumulation piled up in the Inbox. (It was only a week so I haven’t had to plead E-mail bankruptcy). Usually when I’m off on vacation, I try to check in with my mail occasionally, but this trip the closest thing to wireless infrastructure would have been carrier pigeons. Actually after watching the eagles hunting over the lake even carrier pigeons wouldn’t be all that reliable J
The e-mail thing got me thinking about Maximum Uptime verses Guaranteed Data Delivery. Anyone who uses e-mail and leaves the office uses these data concepts without even thinking about it. Yet when people are designing their OPC communication infrastructure these concepts are often neglected or misunderstood.
Maximum Uptime is exactly as it sounds. Your setup is designed to maximize access to the data as much as possible. For e-mail users that means taking a laptop, cell phone or Blackberry along. You data availability depends on your setup. Ad-hoc wireless hotspots, plain text to cell phone, full e-mail access on a hand-held or even satellite connections. Savvy e-mail users choose the right equipment or phone plan to meet their traveling needs. For OPC users maximum uptime typically means redundant communication channels, and the data availability also depends on the setup. The system might use Device level redundancy so each OPC server supports multiple redundant communication channels to underlying devices. It might have Server level redundancy so each OPC client can failover to redundant servers, or it may have redundant OPC client capability. Or any combination of the three. System designers also have to think about how fast failover transfers need to occur and what trade offs in system performance or loading is acceptable to achieve these times.
Guaranteed Data Delivery is a slightly different design. Here the goal is to ensure all the data is captured, but it is acceptable to have delays in accessing the information. Guaranteed Data Delivery is pretty much inherent in any e-mail system. The e-mail server buffers the incoming e-mails until the client downloads them. All the data is there, even days after they hit the server. (Even if you get 183 Unread Inbox Items in a week). For OPC users Guaranteed Data Delivery designs usually incorporate OPC data buffers to store the data, and OPC HDA applications to ensure the data makes it to the final destination.
In practice many mission critical systems the design would incorporate both aspects of redundancy and data buffering.
What about the masses out there? Do you maximize uptime and stay connected on vacation or more rely on guaranteed data delivery and just sort the pile when you get back? What about your OPC systems? Redundancy? Data Buffering? Neither? Both?









