this is the art of network engineering podcast in this podcast we'll explore tools technologies and talented people we aim to bring new information that will expand your skill sets and toolbox and share the stories of fellow network engineers welcome to the art of network engineering i am aj murray at no blinky blinkyblinky and tonight i've got tim at timbertino tim how you doing tim things are going well aj what's new with you this week oh a few things but i can't talk about yet all right and who knows by the time this episode releases it'll be public info so who knows i'll continue to hold my breath yes so will i but hopefully it'll be public soon uh i i love the the time struggle that we have to deal with like we we edit where we record one night and then we edit and then we post and sometimes it's a week later sometimes it's a couple weeks later sometimes it's the next day you never know when it's going to drop plus right now we're recording two times a week to try to build up a cache of stuff so we have plenty of content to get us through the holiday season so uh that makes it extra challenging yeah we'll make it through yep we'll we'll get through and that sound means it's time for the wins for this week winning we have um utc 24 past the ccna all the way from manchester united congratulations here's a good one uh brian from the podcast comptee with your se announced he's starting a new role at cisco as a cyber security tsa or technical solutions architect congratulations brian and if you haven't checked out comptee with your se yet go to your favorite podcatcher subscribe give it a listen you're going to love it andrew roderos held a successful red versus blue event where he built and supported the vms for the event congratulations andrew that's fantastic andy accepted a position as a senior network engineer congratulations andy seems like not that long ago about 11 months actually that you joined the community and accepted a role as a network engineer so i love to see people moving up congratulations mike vest had their contract employer extend the contract another year because they were impressed with his work and work ethic congratulations mike isaiah engineer passed the ccna and is getting some huge kudos on linkedin so congratulations uh david penaloza passed the ccde written uh congratulations david super happy for you i know how long you've been working on this and this was like the last chance to take version 2.0 version 3.0 just came out and is going to be taking effect soon so uh congratulations david love following you and wish you all the best on on the lab for the ccde so congratulations to all our winners this week if you want to be announced on our show you can go to our discord channel and announce it in the winning uh section you can if you're not already a member you can join at artofnetenge.com forward slash iaa tj4 it's all about the journey no uh new patreons to announce this week but if you're interested in joining our patreon uh program you can go to patreon.com forward slash art of netenge and we thank all of our patreons for their support as well as everybody else's support of our show now back to the show tonight we have a guest i'm very excited about if you watch networking field day 26 you may recognize him it is uh stefano he is the ceo of netbeans stefano thank you so much for joining us yeah thank you aj and tim for having me pleasure being here yeah thank you again for for taking the time um i am super excited to be able to share uh your story and netbeans with our audience so so again thank you for taking the time if we could kick it off who are you what do you do yes um my name is stefano gridali i'm a co-founder and ceo at netbeas netbiz is a network performance monitoring um we we tend to monitor the network differently from snmper netflow to go straight into the details we monitor networks from the user perspective because uh for us as former network engineers um it was the best way to detect problems and troubleshoot problems so um as a co-founder i mostly work on product management and um basically work with the sales and marketing team primarily okay so you said former network engineer so is that what you did before starting netbase yeah that is correct i i was a network engineer between technically 2007 until i we started netbeans in 2013. so a good six years was a very great um you know it was the job of my life um but then you know that this happened uh as a reaction for all the struggles uh that i had personalized uh in um you know troubleshooting networks and most probably they were shooting end user tickets uh so so basically like i i i team up with a couple of friends panos empanico and uh we started the company in 2013 um in that year we were kind of in between uh myself i was working full time and then start working on netbeans the concept and then you know panico was finishing his phd and pano was doing his post doc so we all slowly within a year bridge into the actual company full-time as we were able to pick up some you know traction very cool very cool yeah so it was kind of formed out of uh necessity right yeah definitely it was mostly an obsession uh really an obsession with the you know uh there's got to be a better way to to handle this type of problems it's incredible like you know we feel like network monitor was not a network engineer asset that they use much to troubleshoot problems sure you know everybody like jumps into a cli uh telnet ssh to routers and switch and start the troubleshooting process which is very exciting and um i think it's an art however is you know it doesn't scale well and it tends to be very um not there is not a scientific process kind of between around it really in the sense where it's at the moment you don't have all that that you need to actually understand what the root cause could be so we decided to you know in 2013 uh the the market was different than what it is now and so we decided to start the company as an obsession basically yeah that's awesome so as a network engineer what were some of the issues you were seeing that that you saw a gap that you didn't have enough information or or something like that what were some of those issues yeah so i mean the the number one problem was users would detect problem before us and uh generally like especially myself i used to work for a healthcare center so we had like 400 plus remote locations um mostly mpls or you know metro ethernet or other ways to connect remote sites but so remote user would like uh complain about a specific issue and then we would uh you know hold lens of deck when we see not one two three user serving uh probably maybe not connecting to a specific application and um and so the problem that we want to address is like hey why do we have to wait for the users to tell us that the network or suppose is the network having problems right so that was number one and the second was okay now we realized we have a problem now it's all lands on deck uh you know the the typical uh war room conference calls where everybody jumps in and the managers are breathing on your neck to understand what's going on and provide updates and that process generally um was very like um you know random because everybody enjoying a a a phone bridge basically a conference bridge and then try to uh collect data from the remote user understand what exactly they're seeing and then from then there you try to nail down with what the root cause could be um so those were the two major problems that i've seen over and over happening then of course um you know the the root cause could be different things sometimes a bad fiber right sometimes a routing policy that changes your routine and maybe now you have a symmetric routine and you know your voice call web calls are blocked sometimes it's an empty issue because you know someone changes or maybe your um you know uh metro internet carrier changes them to you and nothing works or at least some applications don't work so root cause is different but the pattern is always the same generally like there are kind of softer configuration changes or route policy changes that you cannot really detect with snmp with a traditional network monitoring tool that break your end-to-end connectivity network performance and you cannot detect it out unless you rely on the end users so what we thought is that okay instead of relying on end user why don't we have this um sensor network network of sensors that continuously you know runs end-to-end performance measurement and then we can detect end-to-end uh problems uh you know full service loss of service or performance degradation um so that's what really in 2013 there was there was none at this there were not other you know solution that would do end-to-end um performance measurement at scale um but then also now to add um you know 2012 the raspberry pi came out and now you have a good way to deploy in a cheap way a remote monitoring sensor at scale for you know a few bucks basically and build the solution top of it which that's what we did yeah very cool very cool so what was it what was it like essentially starting a company was that was that an easy decision was that uh difficult was it scary yeah great question yeah right no absolutely um i no it definitely was scary um you know i think you do little by little you commit always a bit more a bit more we don't realize it then you made enough progress that now you can cut your ties with your full full-time job and give it a shot on you know pay yourself and of course getting your salary with the revenue basically right we also did some fundraising but we we didn't really raise a ton right we're really growing organically but very scary but i think for us was easy and somewhat simple because we did little by little so we initially applied to a local startup accelerator here in pittsburgh halls called alpha lab and you know around the us and the war it's full of a startup accelerator so uh that that was the real first you know um team that uh basically invested us like some people that were not the core team of netbeans that like the idea they had a program around the building startups from scratch you know give an idea um so that gives you some traction where okay they help you incorporating the company you know making sure that you can build governance around your company then they give you some funding but nothing crazy you can go full time in three um but you know it was a little progress so it was not as scary as if if you know overnight you decide to start a company and also what we did we we try to you know um first go full time then part-time with our current job and then start putting weekends and evenings on on netbeans and then when you see you have a bit of revenue then you start thinking okay i can go full time right so we did it kind of in a very um careful way so it was not too scary yeah but of course there are some days where you're a bit concerned or you know nervous about your future so there are different uh sensors in the netbeans portfolio there's there's hardware sensors wired and wireless also software-based sensors did you have that full portfolio to get started or did you just have some some basic hardware sensors first what did the beginning look like yeah the beginning was primarily raspberry pi's um in and that was 2013. so then we expanded so think about in 2013 there was not sdying i think started really becoming kind of known i think in 2015 uh at least what i've seen um so in 2013 there was nothing where you could um basically compute at the remote branch office was not a thing right you would have like you know your cisco router or whatever router and switch where you cannot run anything on top you don't have a vm or an environment in the branch office so the only thing that you can deliver an active endpoint or a network sensor is by putting a piece of hardware and that's what we did so 2013 was just pure hardware then in 2015 we added the the wi-fi sensor but was still harder and then in 2015-16 we started uh building more vms um linux packages actually in 2015 also we did a partnership with cumulus network which was a basically router routine operating system based on debian so we did the software agent and then the vm um which could also be deployed especially in the data center as kind of a reflector or a you know a hub sensor in a hub-and-spoke environment where you want to test throughput from branch office to hub so that the vm would cover that role and then we started docker and then we added uh the actual endpoint in response also to the 2020 lockdown for kovid so we're really expanding them the delivery method for our technology but the core right um concept is the same also because it's all based on linux both server and agent so it's very easy to you know repackage the agent in multiple ways so let's talk about what the operations look like at a high level so let's just take an individual hardware raspberry pi type sensor what are some of the constant monitoring tools that it's doing and are there synthetic testing on top that that customers can do active testing as well yeah so um so the agent do the monitoring so different from snmp that's a centralized server you drop somewhere in the data center in the cloud and then they basically do snmp gets and of course more right um but the point is that the monitor is kind of done from from the server perspective in netbeans the monitoring is done only from the remote agents and the service like like a controller like in the wireless right word it's just a controller that aggregates the data and then does its magic um so the agent can run primarily um test uh things like ping dns http and traceroute then we added path analysis uh and those are tests that are run real time so every x seconds a pretty high frequency that is push real time and you see the granular data on the dashboard synthetic wise we do http http as gets we don't do the full http transactions because our primary goal is to prove whether or not the network is working so we all care about what is the network path from a location to a destination um and you know record all metrics such as packet loss latency of course with path analysis now we have traceroute for ecmp network so we can see how the network topology is uh um you know the ecmp topology with the per flow load balancing and then we do of course basic dns and http check because that's very important for the user experience and then we also the ability to run throughput tests basically we instrument uh iperf not sure if you're familiar but it's a tcp or udp basically packet generator um which is open source so we can do tcp or udp throughput test um we do voip similar to cisco psla so you can select the codec g7 11 729 and then you can get the mos core and then we do speed tests um there are different type of speed tests you can run so you can also estimate download and upload speed for you know branch offices or remote users basically what are some of the wireless specific monitoring that your agents can handle yeah so we have a physical sensor uh which generates for on-prem monitoring or for branch office um you you plug basically the sensor reports on the dashboard can be powered via poe or a psu and then we collect we do all active monitoring so we don't do we also now enable packet capture but mostly we basically can do wi-fi connection timing so we connect to a wi-fi network up to four ssid in kind of round robin fashion and then we time the association authentication and the dhcp timing so we can really see how long it takes to connect to an ap authenticate including radius 802.1x for instance as you know you know uh that could definitely cause issues sometime and then dhcp so we make sure that we can get an address um you know and then on top of that we collect wireless metrics such as link quality signal strength noise we collect all the sad bssid information the channel uh pretty much it um for the endpoint for mac um because it gives you also the mcs you can also get the mcs but of course the endpoints are more for home users or you know road warriors versus the the wife and sensor are from prime monitoring so right so you wouldn't put the agent on like every pc or mac in your enterprise right like that kind of defeats the purpose or that that's not the primary intent you would use a remote sensor in in your network and then the agent on on an endpoint outside the network okay and and so talking about the agent i feel like every network engineer or anybody in it when you talk about an agent there's this like instant cringe that you kind of get because it seems like a lot of solutions have an agent and when you layer agent on top of agent on top of agents sooner or later there might be an impact i'm sure the impact on something like this probably isn't very big at all yeah you mean the end point right the user endpoint right yeah yeah our endpoint is very minimal like uh two percent of cpu and about 40 meg of ram um and also we we cap the number of tests uh with the software license that you can run on an endpoint so really you cannot really ex you know oversubscribe to your laptop but yeah it's very minimal impact and it's all overlay so it's not like we're not capturing or listening to what the user is doing which some people freak out we just like run our own tests in parallel and you know just for the wireless metric we don't have the connection timing because we don't want to disconnect reconnect the user from the wireless network but we still capture all the wi-fi metrics that matter to understand wi-fi coverage and performance yeah so to at least understand what the end user's experience is but you're not doing the same test that you might be doing from like your wireless monitoring hardware device right exactly they do the same real-time test ping dns http path analysis throughput iperf um but we just don't reconnect the agent to test the timing like an on-prem physical sensor does that does if the user wants can us configure a reconnect timer just to make sure that basically the network can um authenticate and associate and provide dhcp address to users and now we can also do ssid hopping so i can go from a private network to a public network and then to a private network and in that way i i monitor with the same sensor multiple network and also test the authentication connection timing okay and so from the the endpoint client can you set up some custom tests i mean i'm just kind of thinking like if i'm at enterprises that uses you know office 365 or salesforce can can i see or set up a test that looks at from that user's perspective what the quality of the connection is to those services that i consume yeah so you can um all the tests that we run are user defined so you define a a target or an application by ip fqdn or url so then basically based on that you can you know decide what application you want to monitor microsoft is office 365 is one of the most common um then there could be atlassian products uh sharepoint uh zoom also so yeah a bunch of productivity tools of course sure hey a1 fans aj here to remind you about nordvpn.com nordvpn will help secure you wherever you go i use nordvpn and all my personal devices whenever i'm out about i just go into the nordvpn app hit quick connect and away i go nice and secure don't have to worry about prying eyes anybody looking at my connection if i choose to go work from a coffee shop locally or you know even while i'm traveling if i bring my personal device devices i will use new vpn to help keep things safe and secure i'm using nordvpn right now and there's no degradation in my signal everything looks good when you guys watch us on the live streams so i can't say enough good things about nordvpn they have some great additional services included with their vpn product they'll scour the dark web for your credentials and see if they've been involved in any sort of hacks or anything and then if they have they'll let you know and you can go change your passwords and do whatever you need to do to help keep yourself safe and secure they also have a mode that will block any websites or ads known to possess malware and they just have general ad blocking anyway because you know who wants ads as i record an ad anyway if you want nordvpn and you do go to nordvpn.com t-a-o-n-e for the art of network engineering and you can get a really great deal uh 73 off two years plus four months free so again that's 73 off two years plus four months free of nordvpn if you forget that url just go to nordvpn.com at checkout you can use the promo code t-a-o-n-e for the art of network engineering and we appreciate your support as well as nordvpn's support of the art of network engineering podcast now back to the show so i want to call out something that you said that i think's really important about testing different points in that wireless connection so there's uh the association the authentication the the getting the ip address from dhcp i i like that you have the different tests to be able to show those metrics separately because somebody could complain that hey i'm having an issue getting on the wireless it must be a signal strength issue there must be a wireless issue when there very well could be something with the the dhcp server having latency and handing out ip addresses and it's not wireless at all and to be able to to have those metrics ahead of time i think is really powerful right exactly so i mean the location matters like in a classroom on prem monitoring if you don't drop an end point on the user laptop which you know not everybody wants to do necessarily if they work from all makes sense uh because i think is there is a cost benefit analysis that especially for some companies very very you know um tangible uh but if that you have a bunch of on-prem users that work in the offices um you know a sensor can can do a good job to kind of aggregate be the canary in the mind for that group of users right then you know for sure that you you don't want to put sensor everywhere because that could be over engineering but at least you want to test the basics meaning that my network can associate authenticate and provide dhcp address to the user so that if it works for one sensor will work for the other users then if there is a wireless signal problem or maybe one access point that may be misconfigured you can also detect that if you want um but of course like your monitor is as precise as your configuration and how you configure right so there is always some trade-off to making your monitoring decision where to put the sensor and how aggressive you want to do certain checks right yeah and i want to have like a recommendation for for kind of that the the various deployment scenarios right like i'm just thinking like if you have a remote office with 10 users or 100 users do you just recommend like one percent of users should get the endpoint for monitoring the sensor uh john what i recommend is like uh you can do um you can kind of sprinkle sensor how you feel based on your your um you say um site survey and how your network is laid out uh if that's too much at least i say start with one sensor per remote location especially if it's a small location one per small location at least you you monitor you verify the basics and you know if a user complains they then you already know that hey my dhcp server is working is providing addresses my radius is authenticating right um so um things that maybe the sensor cannot cover because maybe the sensor is in one side of the room and the user is on the other side and then there is not good coverage then you can start straight into the the coverage right or maybe if your sensor is not connected to the same access point that your user that is complaining is connected to you can force the access point the the sensor to connect steer to that access point so there are different options but basically i always say one at least per location if you have a large like a headquarters um deploy the sensor in key areas such as board rooms uh maybe cafeteria whatever like there is a um a congregation of users right or there are very important users in your network that you want to make sure that the service works for them and then could be like actually for example curtis uh he just had a webinar for us in september he works in a large higher education and healthcare system in utah and he for example uses also a dock so he has to support outside the hospital also the dorms and the in the students and he has also kind of a loner type of program where if a student complains in the dorm that the network is not good the water says xyz problem so they can go at the front desk and get the netbeans uh b plug it to their room and then uh the bees collects the data for kurdistan and steam to troubleshoot what the problem could be so long story short some could be permanent and some could be kind of deployed um a doc based on on this on needs but the nice thing is that you don't have to sit there with a laptop for you know two or three days waiting for basically fishing for the problem you just have a remote sensor that you deploy and then gets all the information that you need to to figure out what the problem is yeah i i like that you just you have that that temporary one that floats around when somebody needs it and right you can realize and some time i know it's sometime you prove that the network is right it's not the network and uh you know sometime and that's why also we added packet capture for wi-fi sensors because sometimes to the packet capture and prove that maybe a medical device doesn't follow a specific like um association uh workflow with an access point for instance so um you know sometimes you just have to prove hey everything is good on my end my my sensor can connect so it could be your device of device driver right so right so one thing that i think is really important and helpful about those uh endpoint agents for the people that may be remote where you don't necessarily own the infrastructure or the path to get to them um is is being able to have their information ahead of time and and their device information so i think a pain point for first level support is that information gathering so how do we typically look up a user if it's not a user id it's hey what's your ip address what's your mac address that information isn't always intuitive for uh the end users to get us so out of the netbeas console how how can we look up those endpoints can we look them up by user by computer name what are the options right so now we're talking of course about the endpoint that runs on the laptop right and actually yes um because an interesting thing generally the network agent so we have two type of agents network agents which are the on-prem or hardware virtual sensor and then you have the endpoint right the endpoint came as part of the lockdown and you know work from home kind of shift so and in fact we see now um supporting remote user that work from home is more an l1 help desk type of task versus before the network engine was more a wireless or network engineer tool right so yeah um we basically index um endpoints based on the computer name whether it's a windows or mac uh actually windows is the compute the machine name on mac i think is the username if i'm correct the username of the that installs the the endpoint um but then of course we also log the ap address public and private of the remote user so you have multiple and we also log the mac address right so you have a multiple ways to locate a user um some large organization we work with we also do an integration um and actually network field day our city of panico showed that that part basically we integrate with their own hr or employee management system so when a machine name registered to netbeans you can do the lookup in the employee database and just rename the machine that endpoint to the full name of the of the employee right um so there are different ways that we do it but of course uh when you install the agent general is the machine name for windows which most of the deployments are windows okay so we we've spent quite a bit of time talking about the agents both wired wireless and the endpoint agents how do i see this data where where does it live where is it hosted is there a cloud portal where do i get at it right um so netbiz is a dedicated server per customer meaning that we're flexible with the deploying your netbeans server slash controllers dash dashboard on-prem if you want or in the cloud and if it's the cloud could be hosted by us or hosted by the customer so there is really like flexibility on that is not like a traditional solution where you're in a multi-tenancy environment and um you know uh companies that need that to keep the data on-prem and they have privacy or security concerns they can keep everything on prem uh or in your dmz with fire rules etc so really flexible with that okay so i if we could i'd like to kind of walk through like a you know you you have a problem on your network and you turn to netbeast to help you solve it what what does something that like that look like and i understand that this is a podcast and without doing like a screen share that might be a little bit hard to do but i'd like to try to explore what that might look like in a in a fashion no absolutely i'll start with a very very simple like uh use case um is it the network or is the application right right like a user call from office 223 and says i cannot connect to um you know application xyz now uh well first of all if you have nerves you already know if it's a network an application issue or a user issue right because if it's just one location and you have a sensor there you would see just one sensor triggering a bunch of alerts and then an incident that that agent cannot connect to one or more applications right versus if it's like all the location cannot connect the application then most probably is the application so in that case we understand there is an application issue and we trigger an engineer a target incident right so we call it like global versus global is a problem experienced by multiple users or multiple location or just by one or a subset that is already like a good way that it's a good way to troubleshoot like if you go on starbucks or some coffee shop you cannot connect with the network you you ask someone hey are you able to connect or not right and same journey like when you troubleshoot a problem is this only one user two users all user of the location that's the fundamental question that in some instances it takes a lot of time if you have a large when or a large network or a large user base just answer that question takes a lot of time and then good luck with the communicating with the customer which is going back to the question of team also the power of this is that you don't have to deal and communicate and you know work with end user you don't also distract them you have all the information automatically sent historical information and you can just do your work without needing to interact and disrupt and interrupt someone else's work right right so yeah is it the network of the application that's the number one golden use case then of course if it's the network there are so many things that can go wrong as you know um trying to give tim some room because we've been stepping on each other each other so as you're working through that that kind of troubleshooting workflow if it's the network you know how far up can can you tell there's an issue right are we talking about like internally or like can you see like bgp's lost a peer or or something like that okay what kind of information can we look at yes good question so about the bgpr we don't so think about what an overlay kind of blacks black box monitoring at the moment we don't uh we don't integrate yes an mp or netflow or via maybe a bgp um with the network itself we we see it as an overlay um so the test that i was telling you uh you run those tests from multiple sensors then you have aggregate data you can compare performance from multiple locations from a latency packet loss dns response time http response time and then path analysis with the you can see the topology from multiple point of view and we trigger on changes or performance degradation on all those metrics that we capture yeah um yeah so but we don't do like things like bgp play or we don't see if like i say bgp router lost adjacency or the neighbor right sure um there might be in the future definitely adding you know also uh adding that information to netbeans definitely um but we still have so much to do from this type of natural mountain we're doing that we're not quite ready to also add snmp data or bgp data to the to this solution right right and you know what though like i i appreciate the focus on on the end user perspective right because if if your network is configured correctly but there's still something wrong and the end user is experiencing a problem then you know it really doesn't matter like it's all about the end users performance and their ability to do their jobs right so yeah i appreciate that approach for sure yeah because also so two things of course network engineers want to build networks right learn new technology test those technologies and deploy them and then move on to the next one um it's not nice if you like spend a lot of time then supporting troubleshooting because of course nobody but you knows very well that network and and then guess who they're gonna ping when users have problems and nobody can troubleshoot like level you know the network engineers or architects that build the network uh so we also want a free time of network engineers and make sure that you know um support is enabling power to do troubleshooting at least for the most common and easy problems without having to bug a level two level three right um so that's very important nice nice we're getting some some really good questions in the chat here so a couple of use cases where it might be difficult to get endpoint agents are are manufacturing in you've mentioned healthcare before so how do we get the best visibility there is it is it a combination of uh hardware and wireless hardware endpoints wired wireless endpoints right and i think like uh basically the point is like if you cannot load an engine a sensor on a on a third-party device and we'll understand what's going on with that device why is having network problems right how how can i unders troubleshoot those type of issues yes so our approach is this one like you know um the routing is the same the network is the same so you want to put in parallel a sensor that is in the same say subnet as an x-ray machine or an infusion pump or anything you want to mention and make sure that at least you know for that x-ray machine or in fusion pump to work i'm sure that has to send data and communicate with the server right so i want to make sure at least that the network path between the device and the server is clear and you know there's good performance right um so we basically validate and test the uh network uh from a you know routing perspective and performance perspective that um the network works basically right so you don't need necessary to load an advanced agent on each device and actually initially before kovid we were kind of guess with the building and endpoint although we thought eventually we will do it but that's a big data because when you start loading endpoint in thousands and thousands of laptops good luck getting all the signal but now we have we have an approach to to tame that uh big data problem um right so so that's how you would do it you just put basically the camera in the mind that tells you if the network is good or bad for those devices to function well from a network perspective right then like in the case of uh if you need a packet capture because that device driver is no is and you know it's not up to date which happens a lot with medical devices or industrial devices they have their own stock maybe they have a 10 half duplex interface good luck with that right um there is so much you can do you can just open a ticket with a vendor and see what they can do for you so on the on the on-premises endpoints so if i've got a wired endpoint you mentioned some http get synthetic testing if there's an internal application can i can i do that can i set up uh http gets toward that internal server on a time to basis yes yeah absolutely uh so targets the destination is user defined you you create a target which is an application or a service that you want to monitor and you define it by ip fqdn or url and then you attach to it the test ping dns http tracer or path analysis against that ip so you really test like network connectivity latency dns resolution http http s get uh we can also configure things like um you know you can configure custom http port proxy authentication basic authentication such as ntlm for sharepoint so yeah you're totally um you know you can customize your monitoring based on what you need to monitor and then on top of the you know and then the tests are on pretty high frequency by default the ping test is every five seconds dns every 30 seconds http every 60 seconds and traceroute every two minutes but you can change this value and now you have a time series associated with the performance of that object and then on top of that time series for ping dns http you can run you can attach alert detectors that can alert you if you know there is a loss of being or round-trip time is increasing it crosses a specific sla or just a baseline that you recorded um with netbeas so it's very like uh you know highly customizable um you cannot really see what the tool is doing there is nothing that you know you enable an application you don't know quite what's going on you get sample data no it's like clear in front of your eyes um because we want to make sure that you are fully aware of what the what the solution is doing what is testing and what is the data as it is that then of course on top of the data we can do you know analysis trend analysis correlation aggregation and give you you know extrapolate you know an incident for instance but we also want to give the user to go low level into the details you know of a ping test because sometimes the most difficult problems that we troubleshoot are the one where you know there are a sudden change right so if you have sample data if you have a one minute average of ping test that doesn't help because you know performance may change and you have to see that pattern in your analysis in your troubleshooting are there any existing api integrations if i have other monitoring systems or other systems in general that i want to pull data out of or can netbees go into itsm systems to maybe uh trigger thresholds or open incidents is there any of that available today yeah so we have an api uh to get the data but also we have the i think it's called push api i'm not like a super but basically you can get data from netbeans to uh you know display somewhere else you can also with the api create test or configuration within night b's we also have a way to send alerts or notification via syslog snmp or smtp right so email um snmp traps or syslog and then we have also native integration with splunk pagerduty and slack at the moment um things that we are of course considering web hooks and uh microsoft teams uh are on the roadmap and then eventually service now uh for now with servicenow you can integrate with email which is not optimal but still it works to open tickets when an incident is triggered so definitely like netbeans doesn't we generally exist with an snmp solution a passive monitoring solution it's not like we coexist in a large network monitoring environment and part of the stock so it's very important those integrations are very important so i i'm a deployment engineer for a partner and i gotta ask like what what's a deployment look like so you know someone buys netbees from the the day we start implementing to collecting data what's what's that look like so deploying a typical deployment of netbeas um well it's fairly quick honestly we we sell it as is plug and play but it's really because as network engineer realize you don't have time to configure the server and and of course it's not cloud sas so it's not that you go on a website click the click you create your account right it's on prime but even if it's on prem we basically whether it's on-prem of course if it's cloud is easy it's already there you create your username you plug the agents and then they connect to the dashboard but if it's on prem we work with you we basically the customers submit a ova form where we collect the information about the server ip settings and fqdn and then we build a virtual appliance that they download and install in their environment and then basically if the customer requests hardware sensor they're pre-configured so they can shoot them straight out in the location they plug them in and they report to the server and this is because we do all the server setup ourselves before giving it to the customer so really the customer has to basically load an ova on their system on their you know vmware or hyper v whatever virtual environment they have plug the sensor and they're ready to go if they're softer in the dashboard you have a two-line common script with curl basically that self download install the package and then configure the agent or the vm or the docker to connect to the server so it's really like almost zero touch pro vision provisioning i would say because um we want like a fast deployment we want a faster monitoring so basically once you deploy everything uh which the only problem might be if you deploy sensor hardware right you have to ship them and have someone plug them but if they're virtual with the you know with the cisco integration dna people deploy 80 sensor in like one hour very easily and then of course you have to configure the targets and the tests which that also doesn't take much right um but it depends of course from the use case some have to then tweak a bit maybe the alerts that they want to get um but yeah it could be as quick as a one week or in some cases if you have a lot of sensor maybe taking a bit a lot of more time or if you have like endpoints maybe you want to work with a desktop team or the application delivery team to push it to all the uh endpoints with the sccm or intune or something else okay yeah and so it when you're talking about the endpoints or the sensors do you configure what tests those are doing from within like the the admin panel of of net ps yeah the dashboard basically yeah the sensor is like or at the end point they're dumb they're installed they're there and they sit there and wait for orders and they establish a socket with the server and then the netbeans user administrator configure targets uh and tests select the agents uh that uh where the targets should run and then the agent immediately they start testing um so yeah it's all done through the through the dashboard and that's the main brain and you know i into the network for um yeah for the administrator yeah so so you can customize like across the board like this group of sensors i can have run these tests and this group of sensors exactly run a different set of tests it's not just the same test from all everything right right yeah yeah you could create a agent groups each agent group can have different targets different alert profiles so it's highly customizable and yeah it's not the same test for everyone yeah gotcha gotcha i like that great yeah so i don't want you to give away any uh three to five year you know roadmap secrets or anything but but what's in the uh what's in the immediate future for netbeas well we're actually um pushing out a release next week we we release quarterly and and i gotta say this year was a heavy in terms of in a good way we we published very nice features uh one was path analysis which we demo on networking field day 26 um network thing network networking field day 24 we demo the wireless metric for endpoints and they must estimate for zoom because estimating mass and voip call based on network performance of our zoom is difficult because you cannot deploy an engine there right so we developed that but no uh next week in the short term we're releasing isp grouping and um azure active directory for the dashboard so the user uh they can use azure active directory as sso and then we're going to add more kind of ways to provide sso with the dashboard and um so sp grouping and now i don't remember a pc performance so we can log the cpu memory of the endpoints to see if the user is complaining but the network is all good but we see as an increase in cpu utilization of memory we would see that so at least we can pinpoint remote user experience issues to pc performance uh so those are the three updates again i speak grouping pc performance and single sign-on those are the three updates that are coming next week um and then what we're doing in this um you know this year and next year we're basically like updating our ui um think about the ui that we built in 2013 2014 we did changes along the year but now we're doing a major rework in a react which is a new ui framework which is easier to work with and more kind of user-friendly also from a development perspective so we're doing that so we're updating the reporting functionality we're expanding the reporting functionality to include the information that matters especially when you manage a high number of remote workers so for example isp aggregation information knowing how many users within an sp experience certain issue right or in terms of wireless performance identify users that have bad wireless performance with the report this is right now can a dashboard element but not a formal report so we want to improve that basically wanna have netbeans being used more operationally not just for troubleshooting but also to run you know a help desk with thousands of remote users and then we're also like working on um scaling definitely the solution for thousands of uh sensor of the agents because if you think like now we can scale hundreds of a sensor um and think about this is on-prem it's not you don't have the cloud right uh as a back and you have a vm on prem it needs to scan out to thousands of users which this is what happening now with the work from home instead of having dozens or hundreds of network agents then you have thousands of endpoints to monitor so we're really doing um an upgrade of the back end to make sure that we can scale to thousands of users um tens of thousands of users and then we might also start um also doing more data analysis on top of the data that we collect um so always smarter uh detection of anomalies and correlation so when you release a when you come up with and release a new feature like the pc performance for example is that something that comes from your own research and development or do customers have a way to communicate new features that they'd like to see yeah no definitely so we clearly have a vision we know where we're heading towards a very opinionated vision they we also collect customer feature requests and where the customer votes for each feature i think right now we have about 120 feature requests in our roadmap and uh each some may have one or two votes customer boats some may have 30 customer votes right so we definitely like um we we always like triangle between okay what are the features that have higher votes what are the features that make sense for us to develop right now at the moment based on also where things are heading we try always of course to step to be two-step ahead of what the future will bring right so we make sure that also pc performance for example maybe before kovid we would never even do the end point we thought about doing an endpoint but it was not the top priority then suddenly came because hey everybody's working from home they don't know what's going on with their video performance netbeans needs to help with this use case so it really depends basically on multiple factors but we definitely listen to our customers sometimes we might also you know um we tend not to sell nelby saying okay we'll include this release this feature in the future release but if it is a very important feature that they need and we see that other customer will benefit from the feature we're going to prioritize it right for even one specific customer um yeah but definitely you know we listen to the customer while at the same time with a very a strong vision and a very opinionated roadmap yeah very cool so without getting into like specifics right because i know pricing can be you know very fluctuate between you know is it government is it commercial you know it depends on various things right so yeah what is what does the pricing model i guess look like you know if you buy net b's do you include a certain number of agents or or do you have to like buy agents and packages what does that look like right good question no definitely um so we tend to we want to make the lesson easy and simple to do the math right without a consumption-based model number of tests or different things that are very difficult to then understand what you need how much will cost so we do it basically a number of network sensors that you deployed or endpoint deployed so we have really two classes of agents the network agents and the endpoints with different pricing of course we have a minimal pricing because honestly we sell mostly to medium large size companies um because we're not really so much for smb like maybe this cannot be the only network monitoring solution that you have in your network we always say that if you're looking for snmp then start with the you know other products out there is not a fit but if you already have an snmp a netflow then then you have a pain where supporting user and troubleshooting takes a lot of time and lack of resources then you're in you have the right product in front of you um so you know it's in the thousands of dollars right i would say but it really depends from as as you said the vertical we have discounts for edu of course and uh government nonprofit as well and um but we we work with the minimum number of sensors in terms of network agents or endpoints generally that is about 20 network agents um but you know there could be you know then we're kind of flexible based on the use case and what the customer needs yeah sure cool and if you know anybody listening wants to learn more about netbeans or i don't know if you offer like a free trial or a demo environment you know where can people go to learn more about netbeans yeah thank you so netbiz.net if you've seen the video i have a t-shirt um i don't know if you said straight or like an ambulance you know so netbeans.net um the bees of the network um and yeah we offer a 15-day free trial again because we're not sas we don't have like a self-service free trial right because um during trials we we hosted the dashboard in the cloud because it's easier uh to start a trial and then we have the option if you have software only then we're good to go or if you want hardware you can also request hardware uh for evaluation harder and and then yeah we'll work with you really um throughout the evaluation um you know it's everybody says it's simple easy to use and intuitive but still we work with the with the user the prospect to make sure they understand and we advise them how to use it right um yeah and yeah anybody can go and request a demo and we're very you know responsive so and if not if we have a lot of videos online of course networking field days and you know we have a youtube channel so there is plenty of information to have an idea of how it looks like and we can also give access to a demo dashboard to you know poke around basically awesome very cool well um that hour went by really quick this has been a great conversation tim do you have any last-minute questions and i'm just cruising the the chat here uh no nothing last minute um the questions at least one one of the things i wanted to highlight was a big thing that i've said before a big thing of network field day 26 was uh visibility was a big theme in netbeans was was and is obviously a big big player in that so check them out absolutely absolutely uh stefano i i want to give you an opportunity is there anything that we didn't ask you about netbeans that you want to make sure that you you share with the listeners honestly no i think you know was um sorry if i talked too much sometime clearly so but no no it was really a great session very good questions and uh really appreciate you know us having me here in this podcast it was really great yeah thank you so much for joining us thank you to our patreons you guys asked another great round of questions if you're interested in being an art of network engineering patreon you can go to patreon.com forward slash art of netenge and uh we thank you so much for joining us this evening stefano definitely got to check out netfees and we'll see you next week on another episode of the art of network engineering podcast hey everyone this is aj if you like what you heard today then make sure you subscribe to our podcast and your favorite podcatcher smash that bell icon to get notified of all of our future episodes also follow us on twitter and instagram we are at art of net edge that's art of n-e-t-e-n-g you can also find us on the web at art of network engineering dot com where we post all of our show notes you can read blog articles from the co-hosts and guests and also a lot more news and info from the networking world thanks for listening you