 Hello everyone, my name is Maciek, I will be talking about the Anthropology discovery and visualization and the software I made for that purpose called Wired. As you probably know, topology is the study on space based on graph theory and so the computer network can be presented as a graph. When the nodes represent subnets, we're talking about logical topology. If the nodes represent devices, we're talking about physical topology. Network topology is something to mostly focus on during network design stage, but not only NMS network monitoring systems are pieces of software to keep an eye on the network devices and they often offer various kinds of network maps both automatically and manually created by user. The purpose of map in NMS is to inform all about the location of the failure which can be critical when the failure requires the user to come over and fix it when it cannot be fixed remotely. SNMP is the most commonly utilized source for information when it comes to topology discovery. It works like on the SNMP-enabled device there's a MIP which is the hierarchical database with nameless root and it stores the information about hardware and software configuration and performance of the device. You request the information by the object ID and in return you get the object value. Here are the MIPs which are especially important during the topology discovery. This is the ARP table, this is the table of the network interfaces of the device, their address configurations and of course some description of the system. BridgeMIP is available only on switches and it contains forwarding database table which tells us what device is connected to which port of the switch so based on that we're able to recreate physical segments. Commercially available NMS programs are usually very expensive and designed to work in large environments so they will never show you a single map of the environment. They break the environment to the smaller scopes like routing maps, logical maps of individual subnets and physical segment maps within subnets. This is to avoid the situation where the map is enormously large, doesn't fit to the screen and is hardly readable. There's a simple workaround for such situation. I suggest to add one more dimension so I created that program called wired which scans the network, reads from MIPs, collects the information locally and then it runs, run out. This is another program developed by me which interprets the collected data to produce the three dimensional topology model of the scant environment. Let me show you two examples to present how it really works. Here's the small home network, three laptops, one handheld, one wireless access point, two virtual servers running on the laptop. On this laptop we start the scanner and this is what we get. The red block represents the root of the topology tree which means this is the laptop we started the scanner on. Two leads coming out, one connects to the physical network, the other to the virtual network. As you remember there were two virtual servers, here they are, and this blue box is the virtual hub, the presence of which was assumed by the program after three IP addresses were discovered in the subnet but no switch. So the program created the hub, put it there, connected nodes to it to make the layout correct from the physical point of view. Same thing happened in the physical network because the access point is the router and it does some bridging as well. Two other laptops are here and the handheld. Actually none of these devices were SNMP enabled so the scan was only based on ICMP. Here's another example, this is the industrial network. There are over 70 nodes, several switches, they were SNMP enabled so forwarding tables were fetched from them and successfully processed to recreate the physical segments. Black blocks are routers so as you see the program draws logical segments along horizontal plane and physical segments along vertical plane to save space and to make the routing path between nodes clearly visible. You may be wondering what is it good for? Why would someone need something like that? Well, picture this. The NMS scan your network, produce such models, such scene and gives you the feature which allows you to move the blocks, change the locations, add floors, walls to create the three-dimensional model of your office building, server room and then place the nodes to the rooms where they really are in the reality. No flat map would be more accurate than that. Furthermore, let's say you can interact with the objects in the scene, like you right-click on one of them, menu pops up and you have the options to configure notifications, browse, performance reports, launch scripts or remote sessions so you have the graphic user interface very intuitive and very user-friendly. I'm actually seriously thinking on putting these ideas into practice. Right now I'm working on rewriting the scanner code in C++ because originally it was the per script so the new release is incoming in a couple of days. I'd like to invite testers and coders to join the project. My name is Maciek Kolbursz. You can find my profile at Sourceforge as well as the home sites of my project. Also, have a look at YouTube. There's a short movie presenting Wired working on the LAN. There's some cool music in the background and the movie was made using only command line tools. Okay, I think I'm done. Thank you for your attention and have a good day. Thank you.