Skip to content

Latest commit

 

History

History
163 lines (139 loc) · 6.87 KB

README.md

File metadata and controls

163 lines (139 loc) · 6.87 KB
Raw

Synopsis

NetGrph is an abstract network model for automation, providing a unified network view across diverse network components in order to manage them as a single system via the Neo4j Graph Database. This enables you to navigate your traditional LAN/WAN and/or mixed SDN networks as interconnected nodes and relationships in software, all modeled via the network configurations as they exist rather than secondary sources.

NetGrph can perform universal L2/L3/L4 path traversals, providing context for each layer along the path. It also serves as a VLAN and CIDR database, showing how everything is related. It scales well on even the largest networks, allowing sub-second queries across thousands of network devices. This enables the mapping of complex network relationships for discovery and automation.

Data from queries can be returned as CSV, JSON, YAML, or Ascii tree-art. Network Visualizations can be created by querying the Neo4j webapp as shown below. The data model should translate for use with tools such as D3.js, vis.js or Graphwiz via both the native Neo4j API as well as NetGrph's tree data structure.

All data is accessible via an API, and the lightweight netgrph client can be distributed to multiple machines.

Graph Data Model Example: Vlan 110 -> 200 Traversal

vlan110

[L3 SVIs: Yellow] [L2 VLANs: Green] [Switches/Routers: Blue]

Features

  • Universal Layer2 - Layer4 pathfinding between any two network devices (Full L2 path completion requires NetDB)
  • Path Queries can return a single path, or all ECMPs
  • L3 Network Database of all networks (Automated, VRF aware, and searchable)
  • Search for networks via CIDR or VRF/Role based filters (eg. perim:printers|thinclient, all printers and thin clients in the perim VRF)
  • VLAN Inventory of all VLAN instances across the network, segmented by switch domain
  • Maps L2 VLAN bridges across switch domains, and calculates local/global STP roots
  • Maps L2 paths between devices (regexs supported, eg. dc.* -> dc.* for all datacenter links)
  • Reports the configured VLANs and actual VLANs existing on each link for all L2 paths
  • Secure REST API Server and Client
  • High performance, low latency queries (All queries are sub-second)
  • Easily extendable to support mixed-vendor environnments
  • Ansible playbooks for a five minute install on Ubuntu 14.04/16.04

Planned Features

  • Rewrite API to be more consistent
  • Develop a Web frontend
  • Integrate with Napalm for datasources

Requirements

  • Python 3.4+ (recommend running via virtualenv)
  • Ubuntu or MacOS (should run on any Python compatible platform, but I only support these)
  • Neo4j Graph Database and Java8
  • For Cisco devices, must provide stored configurations (See Rancid / Oxidized)
  • Requires CDP/LLDP Discovery Data. ndcrawl will output this in the proper format for NetGrph with the -ng_file option.
  • Third-party network devices need to be parsed into the NetGrph CSV format
  • Please send me any parsers you create

Documentation

Path Traversal Example

$ ./netgrph.py -p 10.26.72.142 10.34.72.24

┌─[ PATHs L2-L4 ]
│
├── L2 Path : abc7t1sw1 (Gi2/42) -> abc7t1sw1 (Gi1/38)
├── L3 Path : 10.26.72.0/22 -> 10.34.72.0/22
├── L4 Path : VRF:default -> FwutilFW -> VRF:utility
├── Lx Path : 10.26.72.142 -> 10.34.72.24
├── Traversal Type : All Paths
│
├─────[ SRC 10.26.72.142 04bd.88cb.xxxx abc7t1sw1(Gi2/42) [vid:260] ]
│
├───┬─[ L2-PATH abc7t1sw1 -> abcmdf1|abcmdf2 ]
│   │
│   ├─────[ L2-HOP #1 abc7t1sw1(Te5/1) -> abcmdf1(Eth1/8) [pc:1->108] ]
│   │
│   └─────[ L2-HOP #1 abc7t1sw1(Te6/1) -> abcmdf2(Eth1/8) [pc:1->108] ]
│
├─────[ L3GW 10.26.72.0/22 abcmdf1|abcmdf2 ]
│
├───┬─[ L3-PATH 10.26.72.0/22 -> 10.25.11.0/24 ]
│   │
│   ├───┬─[ L3-HOP #1 abcmdf1(10.23.74.11) -> core1(10.23.74.10) [vid:2074] ]
│   │   │
│   │   └─────[ L2-HOP #1 abcmdf1(Eth2/26) -> core1(Eth7/27) ]
│   │
│   ├───┬─[ L3-HOP #1 abcmdf1(10.23.74.21) -> core2(10.23.74.20) [vid:3074] ]
│   │   │
│   │   └─────[ L2-HOP #1 abcmdf1(Eth3/8) -> core2(Eth4/25) ]
│   │
│   ├───┬─[ L3-HOP #1 abcmdf2(10.23.78.11) -> core1(10.23.78.10) [vid:2078] ]
│   │   │
│   │   └─────[ L2-HOP #1 abcmdf2(Eth2/26) -> core1(Eth8/25) ]
│   │
│   └───┬─[ L3-HOP #1 abcmdf2(10.23.78.21) -> core2(10.23.78.20) [vid:3078] ]
│       │
│       └─────[ L2-HOP #1 abcmdf2(Eth3/8) -> core2(Eth8/25) ]
│
├─────[ L4GW 10.25.11.0/24 [rtr: vid:1601 vrf:default] ]
│
├─────[ L4FW FwutilFW ]
│
├─────[ L4GW 10.25.12.0/24 [rtr: vid:1602 vrf:utility] ]
│
├───┬─[ L3-PATH 10.25.12.0/24 -> 10.34.72.0/22 ]
│   │
│   ├───┬─[ L3-HOP #1 core1(10.23.74.10) -> abcmdf1(10.23.74.11) [vid:2461] ]
│   │   │
│   │   └─────[ L2-HOP #1 core1(Eth7/27) -> abcmdf1(Eth2/26) ]
│   │
│   ├───┬─[ L3-HOP #1 core1(10.23.78.10) -> abcmdf2(10.23.78.11) [vid:2462] ]
│   │   │
│   │   └─────[ L2-HOP #1 core1(Eth8/25) -> abcmdf2(Eth2/26) ]
│   │
│   ├───┬─[ L3-HOP #1 core2(10.23.74.20) -> abcmdf1(10.23.74.21) [vid:3461] ]
│   │   │
│   │   └─────[ L2-HOP #1 core2(Eth4/25) -> abcmdf1(Eth3/8) ]
│   │
│   └───┬─[ L3-HOP #1 core2(10.23.78.20) -> abcmdf2(10.23.78.21) [vid:3462] ]
│       │
│       └─────[ L2-HOP #1 core2(Eth8/25) -> abcmdf2(Eth3/8) ]
│
├─────[ L3GW 10.34.72.0/22 abcmdf1|abcmdf2 ]
│
├───┬─[ L2-PATH abcmdf1|abcmdf2 -> abc7t1sw1 ]
│   │
│   ├─────[ L2-HOP #1 abcmdf1(Eth1/8) -> abc7t1sw1(Te5/1) [pc:108->1] ]
│   │
│   └─────[ L2-HOP #1 abcmdf2(Eth1/8) -> abc7t1sw1(Te6/1) [pc:108->1] ]
│
└─────[ DST 10.34.72.24 000a.b004.xxxx abc7t1sw1(Gi1/38) [vid:340] ]

Installation

See the Install Instructions

Support

See refer to the documentation first: NetGrph Read The Docs

You can open an issue via GitHub, or if you would like to speak with me directly, I monitor the #netgrph channel the networktocode slack group. Please try and contact me there for any interactive support.

Contributions

Please see the Contributions document in docs for information about how you can contribute back to NetGrph.

Contributors

License

NetGrph is licensed under the GNU AGPLv3 License.