I’m not going to lie, the UI of pathSolutions TotalView web interface is circa-Windows 95. The old addage, “Form follows function” certainly fits. Now that that’s out of the way, don’t let the simplicity of the interface fool you. The functionality of this troubleshooting software is so well thought out, the interface will fade out of your mind and you’ll just wait for the redesigned look down the road.
PathSolutions presented their suite of performance management tools, known as TotalView, at Network Field Day (#NFD20) in San Jose, CA. Tim Titus (CTO) had spent years in the trenches as a network engineer trying to slog through disjointed alerts and research information during troubleshooting. We’ve all been there at one time or another; checking Google for reasons behind some criptic syslog message or interface counter. This is what made pathSolutions build a toolset that would help bring network errors to the technicians attention, and provide context-based information for the errors with helpful steps to resolve them.
When I mention “context-based” assistance you really have to see it to recognize how great an idea it is. Let’s say your sitting in your cube watching a cat meme, when you see an alert emailed to you from pathSolutions TotalView. As you check on the error you see one port on your core switch is registering interface errors. You make your way over to the TotalView interface, find the port giving that had triggered the alert, and on the right-hand side of the page you see a Wiki of what the problem means and possible troubleshooting steps. In this case, you may be seeing FCS Errors in the alert; the Wiki provides the definition of the problem (official and colloquial), then provides suggestions to resolve the problems most closely related to that error. While that may not sound like a big deal for engineers who have been in the field for a long time, think about how that may help the junior technician or help desk staff who are new to networking. Now they can learn the trade without having to escalate relatively easy problems.
TotalView monitors all interfaces on a network device and polls 19 interface counters to give a complete picture of health. One of the questions that came up at NFD20 was about the polling methodology. TotalView is SNMP-based with most vendors MIBs pre-installed. Engineers have the ability to import custom, or other vendor, MIBs as required. The concern expressed by some delegates is if you have thousands of switches in your environment it would be possible to miss network performance data between polling cycles. This is a common problem for SNMP-based monitoring systems. When asked if pathSolutions was considering moving away from MIB polling in favor of telemetry, the answer may be disappointing albeit reasonable. PathSolutions believes that SNMP provides very rich data points which are lacking in telemetry solutions. While they believe this will change in the fututre pathSolutions have decided to wait until the market shakes out a little bit.
The main question still hung out there, how does TotalView avoid missing data in large environments if it polls at the default of 5 minutes? Their answer is a bit of a differentiator from other monitoring platforms. PathSolutions is single threaded, so rather than send a single request to a 48-port switch to walk the MIB and parse the interface statistics, it sends a single 512k/interface request for the information–much more specific. This not only makes the monitoring lightweight on the network, but it also ensures the rapid and complete collection of performance data. For enterprises with large(r) newtork environments, this collection method could be a game changer especially when combined with the context-based helps.
Not everything is perfect in TotalView-land, though. There were a couple of items that weren’t fully answered. It seems there is limited support for IPv6—which may not be a big deal for most enterprises—and the documentation for IPv6 and other network services is a little sketchy.
One of the items that may stand out to you is the mapping of your network as discovered by TotalView. It’s a great help so you can see where problems are happening. Unfortunately, Tim said the software was not able to export that view into tools like Visio. Of all the imperfections (and there weren’t many that I saw) this was a big one. The reason behind this decision was that most networks are in a fair amount of change and a Visio document only snapshots what “was” rather than what “is”.
I understand that, however, when many engineers work on network projects, they begin with a diagram and make the necessary modifications to it for project documentation. Then it’s all about executing to the documented plan. It seems an easy thing to do to export the information already collected in your monitoring tool rather than start from scratch building your as-is diagram. To be fair, Tim acknowledged that you could export to PDF or other formats, but you’d still have to create the diagram in Visio. I hope this is a feature that gets added even before they decide to tackle the UI -beautification project.
All-in-all, I liked this product for it’s logical layout of errors, resolution path, and light impact on network traffic. We didn’t investigate cost and licensing, but I’m certain the good folks at pathSolutions would be happy to field your call on the subject. I’d recommend looking at pathSolutions for your monitoring requirements…they have a lot to offer.