Accurate monitoring means more effective management of IoT infrastructure
Network monitoring specialist Paessler AG, based in Nuremberg, Germany, and IoT services and low power wide area network (LPWAN) connectivity provider Sigfox, based in Labege, France, announced a partnership agreement in October 2018. The agreement includes the delivery of Paessler’s PRTG Network Monitor for IoT solutions that will enable monitoring data from Sigfox-enabled IT infrastructure sensors and other devices equipped with Sigfox connectivity, which includes Sigfox’s dedicated network that combines its LPWAN and Sigfox Cloud services. Robin Duke-Woolley, the chief executive of Beecham Research, interviewed Martin Hodgson, the country manager for UK & Ireland at Paessler AG, to find out more about the company and its PRTG monitoring platform.
Robin Duke-Woolley: What is PRTG?Martin Hodgson: PRTG stands for Paessler Router Traffic Grapher. It is a monitoring platform now used by more than 200,000 system administrators worldwide. It started out and still is used extensively in the traditional IT space, monitoring both infrastructures, physical and virtual – operating systems and apps – and in effect the whole stack, so that system administrators can see exactly the health and performance of their systems from a single dashboard. Over time we’ve built on that solid foundation to extend it into the emerging requirements of IoT.
RD-W: Do you see the requirements for IoT as different to what you’ve been measuring to in the IT space?
MH: I’d say that they are both not different and completely different. What’s not different is that there’s still a requirement for an insight into the health of the devices, their ability to deliver whatever that function is, but what’s different is they require an entirely different set of tools and techniques.
RD-W: What do you see as the main challenges of dealing with IoT?
MH: It mirrors where IT monitoring was maybe 15 or 20 years ago, where you had to use a vendor’s monitoring technology for a vendor’s product. If you had a HP server, you had to buy HP hardware and software to monitor it – and the same again for each of the brands you had, so you ended up with multiple monitoring solutions for your entire IT infrastructure. Our approach is to tap into the available interfaces. PRTG is completely agentless; we use standard monitoring techniques and protocols to extract the information from these devices and processes. We use the same technique with IoT as we have with traditional IT and that ability to tap into vendor siloes such that you have a unified solution for your whole stack.
RD-W: So that’s the similarity. Where would you see it as being completely different with IoT?
MH: In the IT world, we’re used to the concept of always on – our expectation is that all our devices are always on – but in the world of IoT there are many very good reasons why devices aren’t on. Whether we’re looking at something that is currently in flight that doesn’t transmit or whether it’s for power consumption or bandwidth consumption reasons, they are not on-line. So that is one of the things that is significantly different in the IoT world.
RD-W: How long has PRTG been active in the market?
MH: This will be Paessler’s 22nd year. PRTG has existed for fifteen years and evolved over that time. It started out as a dedicated network monitor. Our founding father, Dirk Paessler, tried out a number of techniques and potential products and as part of one of the projects he needed a network monitor and he couldn’t find a good one, so he built one. That’s what became PRTG and that’s still the cornerstone of what we do.
RD-W: The data you extract is still real-time information, in the same sense that you were doing in IT days?
MH: It certainly can be. Monitoring can be wildly different based on the particular use case and that’s where the administrator is alerted by whatever levels are appropriate for his environment. The administrator, or user, sets thresholds. We can also specify at what point to escalate beyond those and use a different method of alerting at the point of escalation. You may start by sending an email to the network administrator when there’s a problem with the network, but then you might send a text message to a manager if that hasn’t been resolved in a given timeframe.
RD-W: How does PRTG scale?
MH: You can start out with a smaller instance and grow over time or start with a larger instance; you can also combine multiple instances of our product into a single pane of glass – two ways you can get to a very, very large scale. We have a free version of our product with the same features we sell to customers, but you can only connect up to 100 sensors or things on that. Beyond the free version the number of licensed elements, or bundles of sensors, is in theory unlimited.
RD-W: What other unique features does PRTG include?
MH: We have our own database technology. We designed a database for PRTG that gives us a huge range of advantages. It means that we’re not compromised by having to fit our tables into a general purpose database, so our data is stored in its original form. While it’s completely user configurable based on available storage and performance, for most of our customers we store the data in its original form for up to 12 months. There are some sensor types that are very data rich which we tend not to store for such a long period and a good example would be flow information – management information from a switch might not be something that you need to keep for such a long period.
RD-W: Does that mean you can store huge amounts of data but you can actually get to that data and sort that data quite quickly?
MH: Yes – absolutely. We designed our database for monitoring. Most of our competition in the traditional monitoring space use general purpose databases.
RD-W: What advantage does that give you?
MH: We would say that because it’s specific for purpose, we get better performance from it. It also means that the end-user of the system doesn’t need to buy something else or employ a database administrator (DBA), because it’s self-tuning.
RD-W: Does this enable lower cost or more responsiveness?
MH: Yes of course lower cost, plus quicker time in terms of the installation. PRTG is up and working within about 10-15 minutes but by having a simpler stack and consistency there’s less for the administrator of the system to deal with – it’s one less thing to patch, one less thing to go wrong. Also, it’s built for purpose. If you think about how a database could be used in a wide variety of different remits and has to cope with variable table size, number of records and other factors, we know that in monitoring there are going to be tiny fragments of data and there’s going to be lots of them.