What is our primary use case?
We were originally using NetBrain for mapping and that was our primary purpose. Since then, I have developed a couple of unique functions inside NetBrain that go out across the network and locate specific pieces of equipment. The parameters for that equipment is a very unique state of interface, a switch port interface, drawing on POE power. The devices I'm looking for are Primex clocks and they continue pulling POE power, but they shut down the interface for data. It creates a down interface state, but a POE up state which is very unusual. I use NetBrain to crawl through the site, particularly where we had a bunch of these installed and were looking for those two conditions. It was a matter of pulling and comparing to get a final result value. That was very helpful. I'm a senior systems engineer and we are customers of NetBrain.
What is most valuable?
The original case was mapping and the solution enabled us to draw out our maps instead of hand draw the maps with the links and everything else. It's okay on the value, a bit expensive for just maps, but makes it a lot easier to see things visually and take it from there.
What needs improvement?
Creating the parsers and QFs from model to model show that the IP interface brief isn't consistent. From a 6500 to a 3750 to a 9300 Cisco switch models, the show IP interface brief is different. There's extra columns and you have to parse for those. Setting those up is a bit of a challenge on this particular version. From what I've been able to gather, the next version, 8.X, is better at that and has more of those, but my perception is that building the parsers out is a bit of a challenge. I've done it but it's a challenge.
Then there's a similar function again with the POE that didn't have the show power in line command and I had to figure that out. Again, that table comes out in a very unique state and then requires parsing down and pulling that out. It's basically using a CLI scraping which should be simplified and a more intuitive process. I don't know if the newer version does APIs or not and if that would have been any easier or better to manipulate, I have no idea.
For how long have I used the solution?
I've been using this solution for a year.
What do I think about the stability of the solution?
So far it's been fairly stable for our needs. It's not like it's hammering the network every day to pull data. It's not like we've got a tied into service now for when there's an incident, we can pull it. We've got the service now. API hooks between the two. It took us a while to figure out what we wanted to do and then we finally got the APIs in place, but we're not utilizing that at all.
What do I think about the scalability of the solution?
We initially bought 250 nodes. We added another 120 some nodes so we need to up our license count because we're nowhere near pushing scalability. I mean, we're a decent sized enterprise, but we've got a magnitude of 500 total individual switches and a total of 150, 200 nodes at the time. Scalability hasn't been a problem.
How are customer service and technical support?
We did find a code bug and reported it and ended up with one of the original developers on the phone with an interpreter because he was Taiwanese, I think, and didn't speak very good English. They took some data and figured out what the problem was, and gave us a patch. The struggle that I have with their tech support is that they don't have American 800 service, it's primary email and they get back to you. It's usually evenings out of West Coast or Asia early in the morning so the timing of some of their responses doesn't always work but It's been okay. Their documentation is pretty good. I was able to go through it and figure out what I needed to do and how to do it. I opened a couple of services when I was trying to do those parsers and build those out.
Because of my programming background, I was able to get through it, and because I've got that mindset I was able to figure out what I needed to get done. When you do need the tech support, it's email and reply back and if you don't specify exactly, or if there was something lost in translation, it's a challenge.
How was the initial setup?
We had the company implement for us, it's part of their professional services. I don't remember if we paid for it or if it was just a part of the package. The polling engine and then the rendering engine, they use a couple of different servers in the back end and it's all Linux-based, and I'm not as strong on Linux or Windows. I'm a network guy.
What's my experience with pricing, setup cost, and licensing?
The licensing is a node-based structure together with a portion of the purchase price. A node is basically an IP address for either a device or a switch stack. We've got some switch stacks that have six or eight switches in them, but because it's all under one IP address, a single management interface was a node. Even though we've got 150 plus nodes, and even though we've got 500 switches, it's because of the way the stacks work. Licensing is on an annual per node basis because the solution is on-prem. Eventually if they push it to the cloud, it will probably switch to a monthly cost. It's more for getting the updates and contacting tech support.
What other advice do I have?
If someone is fairly good with Linux, they could probably get this up and running themselves. Once it's up and running and you don't have the strange requirements I had, and you just want some basic maps, it'll work. The new version, 8.X, has got other features and tying and capabilities for more advanced things. So they're always improving. They're into a niche market, so it works for their needs. They're coming at it from a mapping perspective and then expanding back into automation since they've got the database and resources.
I have nothing else to compare this to other than hand drawing and we all know how much that sucks, so I would rate this solution a five out of 10.
Which deployment model are you using for this solution?