Original Summary

Hi everyone, I'm Cosmin. I work in a NOC in Romania, and I've built NOVA Network Lab : a desktop environment for network configuration and troubleshooting, with real appliance operating systems, an interactive 3D datacenter, multi-site WAN and MCP automation. The core lab workflow, including MCP automation and global failover, is implemented and working. GNS3 was the reference point for the kind of tool I wanted to build. NOVA brings the network and the infrastructure around it into the same workspace: configure a router, trace the traffic, work on the cabling, then explore the datacenter that houses it. NOVA has its own QEMU orchestration and Ethernet fabric . The guests run their actual operating systems, and the fabric forwards real Ethernet frames between them. Each device gets its own working disk while the original appliance image stays separate. Here's what it includes: Real appliances and consoles. Cisco IOSv and IOSvL2, FortiGate, VPCS and graphical Linux guests. Consoles can float, share tabs or open in separate native windows, with saved labs and editable VM resources. Linux desktops without VirtualBox. Boot Linux and Kali guests directly inside NOVA through its own QEMU integration, with graphical consoles and lab networking. Linux guests can also access the Internet through the host's Wi-Fi or Ethernet connection. A connected 2D topology and 3D datacenter. Both views use the same guest devices, interfaces and network links. The room includes racks, rear-panel connections, patch panels, transceivers and exact PDU outlets. Cable faults that affect actual traffic. The RJ45 workbench lets you strip the jacket, arrange T568A/B conductors, seat the wires, crimp and test. Applying a fault or repair changes the real virtual link's carrier and forwarding, so you can see the effect in the guest's ping. Multi-site WAN and global failover. Run several labs concurrently, place sites across cities and countries on an offline world map, and connect them over WAN links. Delay, jitter, packet loss, bandwidth limits and fibre distance affect actual guest traffic. Cut, black-hole or flap a primary link, watch BFD detect the failure and OSPF reroute traffic over the backup, then inspect the recovery. This is failover between network sites, with real router protocols and traffic. Power, cooling and physical incident models. UPS battery runtime, transfers, overload and load shedding; utility outages and PDU feeds; cooling failures, containment and fire-suppression scenarios. The room also has a buoyancy-aware Navier–Stokes airflow solver and thermal maps. These are simulated training models, with estimated power and temperatures. A built-in “mini Wireshark.” Capture real traffic from lab links and guest interfaces, including Linux Internet traffic routed through the host's Wi-Fi. Filter by protocol, search packets, inspect addresses and ports, read plain-language explanations, and export the capture as a PCAP. There is also ana

  • 情报分类：服务器与云资源
  • 分类依据：内容涉及服务器、云资源或网络线路
  • 信息来源：Reddit · SideProject
  • 发布时间：2026/10/14 09:17:40