"Illuminating the shadows of the Internet" Opportunistic network and web measurement Home •• Project home page Overview •• Brief overview and news FAQ •• Frequently-asked questions Pubs •• Publications and people Stats •• Preliminary results Maps Proxies •• Data coverage and proxy maps Contribute •• Join our community Network statistics This page presents preliminary results we've collected from analyzing the traces taken from the collection locations. Overview Statistic Count Unique hosts measured 6957282 Unique hosts running Java 1126168     Unique public IPs 6419071 Percentage IPs behind NATs 73.81%     DNS servers 129670 Prefixes 85048 Unique ASNs 14567 IPs with reverse DNS entries 4577421 (71.31%)     Total countries 214 Unique GPS locations (from Quova) 15490 Unique hosts measured: Number of unique (public IP, internal IP) pairs we've measured DNS servers: Number of unique DNS server IPs identified Prefixes: Number of unique prefixes measured. Prefixes are collected from public BGP data. All geographic location information has been kindly provided by Quova. Requests per Browser type Requests per operating system type (SYN fingerprint) Requests per operating system type (SYN fingerprint/User-Agent) User-Agent and SYN fingerprint for clients with proxy headers (* Proxies skew the relative numbers of OS types when comparing the syn fingerprints and user-agents. In particular, the relative number of Linux and BSD machines is much greater given the SYN fingerprint.) NATting per top-level domain NATting per geolocated country (* The NAT rate for the .cn tld is low due to a heavy bias towards educational institutions. Of the greater than 200,000 requests we've received from China, roughly 600 have reverse DNS entries, and of those virtually all are from .edu.cn.) Distance b/w clients and their DNS resolvers (Using clients' external IP addrs)

Geographic distance	Topological distance

Distance b/w clients and their DNS resolvers
(Restricting each line to those clients having topological divergence = k)

All such client/DNS pairs	Include only pairs where traceroute reached destinations

Distance b/w clients and their proxies
(Only clients with publicly-routable internal addrs)

Geographic distance	Topological distance

Distance between proxied clients and server
Comparing direct route vs. detour route via proxy

DNS servers per Public IP	Internal IPs per Public IP

Use of Reserved Addresses behind NATs


Histogram of IPs in 192.168/16
Image is huge, click HERE

Allocation of 192.168.0/24
(192.168.0.2 is the start of the default DHCP range used by smae Netgear home routers)

Allocation of 192.168.1/24
(192.168.1.100 is the start of the default DHCP range used by various Linksys home routers)

Allocation of 192.168.100/24


Allocation of 169.254/16


Histogram of IPs in 10/8
Image is huge, click HERE

Allocation of 10.0/16


Allocation of 10.0.0/8
(* Note: 10.0.0/24 is used as the DHCP range in documentation for the Cisco SOHO77. 10.0.0.150 is commonly used for local interfaces terminating PPTP tunnels and is the start of the default DHCP range for Alcatel SpeedTouch Home routers.)

Allocation of 10.0.1/8
(* Note: 10.0.1/24 is the default DHCP range used by, the Cisco uBR924 cable modem, Apple's Airport Base Station and Aiport extreme, and is used by linux DHCP config files)

Histogram of IPs in 172/8
Image is huge, click HERE

Histogram of IPs in 172.16/16
Image is huge, click HERE

Allocation of 17.16.1/24
(* Note: 172.16.1.33 - 172.16.1.250 is a standard range for DHCP allocation, used by 2Wire's HomePortal http://www.2wire.com/)

---