Platforms Comparison

This section compares nxthdr with other popular platforms for Internet research. nxthdr provides two integrated platforms: Saimiris for active measurements and PeerLab for BGP and routing experiments. The comparison focuses on key features, data access, and capabilities to help you choose the right platform for your research needs.

Please note that information about other platforms is based on publicly available documentation and may not reflect the latest changes. Always refer to the official documentation of each platform for the most accurate and up-to-date information. Feel free to open an issue if you find any discrepancies or have questions.

What Makes nxthdr Unique

nxthdr is the only platform that integrates both active measurements and BGP experimentation in a single, cohesive environment:

Integrated Platform Advantage

Unlike other platforms that focus on either measurements OR BGP experiments, nxthdr provides both Saimiris (probing) and PeerLab (peering) under one roof. This integration enables:

• Joint control and data plane studies: Announce your own prefixes via PeerLab and immediately measure how routing changes affect reachability and latency using Saimiris
• End-to-end routing research: Study the complete picture from BGP announcements to actual packet delivery
• Unified data access: All BGP data, traffic data, and measurement results available in one place under PDDL
• Seamless workflows: No need to coordinate between separate platforms or merge datasets from different sources

Platform-Specific Advantages

Saimiris (Probing Platform)

• Anycast measurement support (unique capability)
• Full control over probe parameters and source addressing
• Framework for measurement tool development

PeerLab (Peering Platform)

• Lease dedicated IPv6 prefixes
• Automated Docker-based setup
• Real BGP experiments on production Internet

Both Platforms

• Truly open data with no authentication, delays, or licensing restrictions (PDDL)
• No application or approval process required
• Immediate access to both platforms

Peering Platform Comparison

This section compares PeerLab with the PEERING testbed, the leading platform for BGP experimentation.

BGP Experimentation Capabilities

Feature	PEERING Testbed	nxthdr (PeerLab)
Access requirements	Application required; academic use only	Open signup, no approval needed
Prefix allocation	Prefix pool or bring your own	Prefix pool as a service or bring your own
Source ASNs	7 ASNs	1 ASN (AS215011)
IXP locations	5 IXPs worldwide (NA, EU, SA) + Universities + Vultr Cloud	4 IXPs in EU + Optional Vultr Cloud
Connection method	Remote peering via tunnels	Remote peering via Headscale overlay
Setup complexity	Manual BGP configuration required	Automated via Docker (PeerLab)
BGP data access	Limited; primarily for your own announcements	Full BGP feeds and traffic data (sFlow) from AS215011
Integrated measurements	No (separate platform needed)	Yes (combined with Saimiris)
Data licensing	Varies by use; research-focused	Public domain (PDDL)
IPv4/IPv6 support	Both IPv4 and IPv6	IPv6 only (currently)

Advantages Summary

PEERING Testbed is better if you need:

• IXP presence across multiple continents
• Multiple source ASNs (7 ASNs)
• Both IPv4 and IPv6 support
• Established platform with long operational history
• Large academic collaboration network

nxthdr (PeerLab) is better if you need:

• No application required: immediate access for anyone
• Prefix leasing as a service to lease IPv6 prefixes directly from the platform
• Automated setup via Docker, Headscale and BIRD for quick experimentation
• Full BGP and traffic data from AS215011 made publicly available
• Integrated with Saimiris for joint peering and measurement studies
• Truly open data with no licensing restrictions (PDDL)

Probing Platform Comparison

User-Defined Measurements

Feature	CAIDA Ark	RIPE Atlas	NLNOG Ring	nxthdr (Saimiris)
Access requirements	Limited to vetted researchers	Open with daily quotas	Requires participation (host a node)	Open signup, no approval needed
Daily quotas	N/A (vetted access)	~100k results/day	No enforced quotas	10k results/day (currently)
Supported protocols	Scamper (ping, traceroute, HTTP, DNS)	Ping, traceroute, DNS, HTTP, NTP	Any Linux CLI tool	ICMP/UDP probes
Source locations	~220 cities / 70 countries	~1,000+ cities / 187 countries	~300 cities / 60 countries	2 cities / 2 countries (expanding to ~40)
Source ASes	~200	~4,000	~547	1 (AS215011)
IPv4 / IPv6 support	Both	Both	Both	IPv6 only (currently)
Anycast measurement support	No	No	No	Yes
BGP joint experiments	No	No	No	Yes (via PeerLab and Saimiris)

Measurement Framework Comparison

PKTLAB is a measurement framework (not a platform) that can be compared with Saimiris in terms of probe input capabilities:

Feature	PKTLAB Framework	nxthdr (Saimiris + Prowl)
Type	Framework (self-hosted)	Platform as a service
Probe specification	Custom probe definitions	Caracal probe specification
Infrastructure	Bring your own servers	Managed infrastructure (AS215011)
Setup complexity	Requires deployment and configuration	Sign up and start measuring
Data storage	Self-managed	Managed ClickHouse database
Data access	Local only	Public API, no authentication

Advantages Summary

RIPE Atlas is better if you need:

• Extensive geographic coverage (1,000+ locations worldwide)
• IPv4 measurements
• Diverse source ASes (4,000+)
• Established, mature platform with long-term stability
• Higher daily measurement quotas if you have credits

CAIDA Ark is better if you need:

• Long-term curated datasets
• Historical data archives
• Established research datasets for reproducibility
• Vetted research environment

NLNOG Ring is better if you need:

• Full shell access to measurement nodes
• Ability to run any Linux CLI tool
• No protocol restrictions
• Direct server access

M-Lab is better if you need:

• Long-term retention of performance data
• Large-scale speedtest and NDT data from end users
• TCP-based performance measurements

nxthdr (Saimiris) is better if you need:

• Anycast measurement support (unique capability)
• No application required: immediate access for anyone
• Integration with BGP experiments (PeerLab) for joint studies
• Full control over probe parameters and source addressing
• Truly open data with no authentication or licensing restrictions (PDDL)
• All measurement data publicly available

Datasets

Feature	CAIDA Ark	RIPE Atlas	M‑Lab	nxthdr
Data access	Public datasets; raw data requires request; delayed availability	Public API; private data with auth	Public via BigQuery (requires GCP account)	Direct ClickHouse access, no authentication
Data licensing	Academic, non-commercial use only	Open access under RIPE NCC terms	Public domain (CC0)	Public domain (PDDL)
Data retention	Long-term archives; curated datasets	Private data ~1 month; public archives	Indefinite retention; 24h publication delay	7 days (rolling window)
BGP data	Yes (RouteViews, BGPStream)	Limited	No	Yes (full BGP feeds from AS215011)
Traffic data	No	No	Yes (NDT, speedtest)	Yes (sFlow from AS215011)

Sources

• CAIDA Ark

  • https://www.caida.org/projects/ark/
  • https://www.caida.org/data/overview/
  • https://www.caida.org/about/legal/

• RIPE Atlas

  • https://atlas.ripe.net/
  • https://atlas.ripe.net/docs/api/
  • https://www.ripe.net/manage-ips-and-asns/community/active-measurements/ripe-atlas/data-access
  • https://www.ripe.net/legal/terms-conditions

• NLNOG Ring

  • https://ring.nlnog.net
  • https://ring.nlnog.net/documentation/
  • https://ring.nlnog.net/users/faq/

• M-Lab

  • https://www.measurementlab.net/data/
  • https://www.measurementlab.net/mlab-scope/
  • https://www.measurementlab.net/privacy/
  • https://www.measurementlab.net/tools/
  • https://console.cloud.google.com/marketplace/details/measurement-lab/ndt
  • https://github.com/m-lab

• PEERING Testbed

  • https://peering.ee.columbia.edu/
  • https://peering.ee.columbia.edu/tutorial/
  • https://peering.ee.columbia.edu/contact/

• PKTLAB

  • https://packetlab.github.io/
  • https://github.com/packetlab