IMCA marine technical adviser Graeme Reid explains how network ‘storms’ can impact DP operations
Over the past year, when using Teams or Zoom for meetings, you’ve probably encountered non-critical minor ‘network storms’ evidenced by that moment when someone freezes. While annoying this is usually harmless and even amusing. Now imagine a ‘storm’ disrupts operations on a DP vessel controlled entirely by networks while a diver is on the seabed, or the vessel is alongside a platform with the wind picking up. A storm in these situations would be anything but harmless or amusing.
First, let’s define a network storm: it is the accumulation of traffic on a computer network. Extreme amounts of traffic constituting a ‘storm’ consume sufficient resources rendering the network unable to transport normal traffic.
IMCA’s DP committee hosted a workshop in early March on DP system network storms attended by vessel owners and operators, third-party consultants, network specialists and OEMs from member companies. Three IMCA presenters were all heavily involved in the subject; and workshop sessions enabled everyone to discuss and debate many specific points.
Above all, we wanted to unravel the mysteries of network storms to ensure that our industry has the knowledge and tools available to ensure DP vessels are robust in terms of failure in this area; and any testing undertaken was fit for purpose and associated risks understood.
The focus was on the complete DP system, not just the DP computers and associated networks, but all systems and their respective networks, from the position reference systems to the propellers and everything in between.
“The entire industry needs to come to terms with this complex and variable topic”
We were keen to address the belief that many network storms currently go unnoticed, being put down as a ‘glitch’, ‘anomaly’ or ‘doubtful human action’. Discussion at the workshop clarified that, although the effects may appear suddenly, they can build up over time and may materialise in many forms, such as crashed controllers; controllers becoming slow; erratic behaviour; equipment freezing or dropping offline.
One of our presenters provided a video of a network storm test being undertaken; it demonstrated complete loss of all thrusters from automatic DP control, coupled with a loss of independent joystick on a DP-equipment class 2 vessel. Throughout the test there was a cacophony of noise from constantly sounding alarms and a host of failures were appearing on the status screen; the combination made it impossible to calmly figure out where the problem lay.
The workshop concluded with arguably more questions than answers. I then took the challenge to the OSJ DP conference in March. Again, this confirmed our belief that the risk of network storms is poorly understood in the DP world. Donald Rumsfeld and his Known-Knowns, Known-Unknowns, Unknown-Knowns and Unknown-Unknowns has become legendary; the DP world fits into the last category, for which read ‘Unknown Risks’. We need to change this; indeed, the entire industry needs to come to terms with this complex and variable topic.
IMCA’s network storm agenda covers four key areas. We need to increase knowledge within system integrators, OEMs (the providers of networks for the overall DP system – not just the DP control) and more generally. We must explore the testing requirements and/or demands on vessel owners and/or operators; currently these appear to have little technical basis.
Producing guidance on the subject is essential and we have already formed a knowledgeable workgroup to take on this task. And lastly, we will be arranging a series of learning opportunities around the subject in all parts of the industry to turn those Unknown-Unknowns to Known-Knowns, ensuring competence lies at the heart of the DP world.