- index-editorial.html: floating \$_ terminal button (bottom-right)
- macOS-style title bar (traffic light dots), backdrop blur 18px
- Guided wizard: category → message → name → submit
- POST /api/feedback with ASN context auto-filled
- Safe DOM output builder (no innerHTML on user data)
- server.js: feedback API endpoints
- POST /api/feedback — stores entries to feedback.json
- GET /api/feedback?token=... — admin read (token-protected)
- OPTIONS preflight for CORS
- FEEDBACK_TOKEN + FEEDBACK_FILE constants from .env
- Host routing: shell.peercortex.org → shell.html
- public/shell.html: full-screen admin terminal
- login command → token auth via API
- list / list [category] — tabular overview
- show <n> — full entry detail
- stats — bar chart by category + top ASNs
- export — JSON file download
- refresh, logout, clear, help
- Upgrade from Leaflet to MapLibre GL JS 4.7.1 with OpenFreeMap dark base
- Add submarine cable layer (TeleGeography via /api/submarine-cables proxy, 24h cache)
- Add global datacenter layer (PeeringDB all facilities via /api/global-infra proxy)
- Layer toggles: ASN PoPs | Submarine Cables | Global Datacenters
- Dark-themed popup styling matching PeerCortex UI
- Server-side caching for both new data sources (24h TTL)
Root cause of 2.7GB RAM usage and 20+ OOM restart loops:
- server loaded all 825k ROAs from Cloudflare RPKI feed into a JS Map
- Every 10min refresh caused double-memory spike (old + new data) -> OOM kill
Solution:
- Remove rpkiRoaIndex Map, addRoaToIndex(), validateRPKILocal(), ipv4ToInt()
- fetchRpkiAspaFeed() now only loads ASPA objects (~1484, negligible RAM)
- Add validateRPKIWithCache(): calls RIPE Stat API per-prefix with a
5000-entry LRU cache (6h TTL) — same API already used by fetchRPKIPerPrefix()
- Update all 4 call sites: sync .map() -> await Promise.all()
Result: 2.7GB -> ~96MB RAM, no more OOM restarts
new URL() throws ERR_INVALID_URL on malformed inputs like XSS probe
requests (e.g. //brusEYkk%22%3E%3Cscript%3E...). Uncaught exception
caused memory leak and process restarts. Return HTTP 400 instead.
server.js: fetchPeeringDBWithRetry now does 3 attempts with exponential
backoff (2s, 5s) instead of 1 retry at 1.5s. Under audit load (9+
concurrent PDB requests), the longer delays let rate limits clear.
audit.py: stagger ASN submissions by 2s so PeerCortex's internal PDB
requests don't all fire simultaneously. Nightly audit takes ~8min
instead of 5min — acceptable for a midnight cron job.
When the same field fails 2+ consecutive audit runs, a known_issue
entry is written into the ASN's registry profile with:
- field name, description of what's wrong
- first_seen / last_seen dates, occurrence count
- last auth vs PC values
- status: open (stays until PeerCortex data matches)
Report shows KNOWN ISSUES section (all open issues across registry).
Issues auto-resolve when the ASN passes, or partially resolve when
individual fields are fixed. Also stores ASN name in registry.
- Add fetchPeeringDBWithRetry: 1 retry with 1.5s delay on null response
- Add fetchJSONWithRetry: 1 retry for RIPE Stat prefixes + neighbours
- Log HTTP 429 from PeeringDB instead of silently swallowing it
- Add &limit=1000 to netixlan/netfac queries (prevents truncation at 250)
- Fall back to asn= / local_asn= queries when PeeringDB net lookup fails
(previously: netId=null → IX=0, fac=0 for ~22 ASNs)
bgp.he.net scraper:
- Fixed prefix regex: "Prefixes Originated (v4): 147" format
- Fixed peer regex: "BGP Peers Observed (all): 274" format
- Added prefixes_all field
- AS6830: v4=147, v6=9, peers=274 (was all unavailable)
- Prefix cross-check now works: RIPE 151 vs HE 156 = 97% agreement
Peering Recommendations:
- Now filters out already-established peering sessions
- 3 categories: New Opportunities, Already Peering, No Shared IXP
- Uses BGP neighbour data to detect existing sessions
- Shows "Already peering with all top networks" when applicable
Dashboard: Added "Score Breakdown — Why X/100?" section showing:
- Per-check weight, earned points, and reason
- Total calculation with formula explanation
- Data source attribution
- "info" status excluded from scoring (e.g. MANRS API auth)
Security: try-catch around new URL() parser — malformed URLs from
scanner bots (XSS attempts) now return 400 instead of crashing server.
Was causing repeated crashes from automated vulnerability scanners.
- Route Server: threshold lowered from 20 to 10 IX for "bilateral policy" pass.
3-9 IX without RS = "info" (not warning). <3 IX = warning.
AS212635: 19 IX → pass (was warning)
- rDNS: sample size increased from 5 to min(20, total_prefixes)
Better coverage for large networks (AS13335: was 5/5621 = 0.09%)
- IX Route Server: always use asn= query (more reliable than net_id when PDB rate-limits)
AS212635: 0 → 19 IX connections correctly detected
AS212635 score: 98 → 100/100
- Geolocation: global networks (5+ facility countries) now get pass
even when MaxMind has no data (was warning)
- Route Server: uses ASN fallback when PeeringDB net_id unavailable
(was showing "0 IX connections" due to rate limiting)
- IX geocode fallback: CITY_COORDS map + IX_CITY_MAP for 70+ cities
AS49544 (i3D.net/Ubisoft): 100 IX connections correctly detected,
bilateral peering policy recognized, 27-country global presence pass
renderNetworkMap() was missing its closing } after the setTimeout(50)
callback. This caused a SyntaxError that prevented the entire script
from parsing — doLookup was undefined, Lookup button did nothing.
Also added deploy.sh backup script on Erik (auto-backup before restart,
keeps last 20 versions of server.js + index.html).
IXPs without PeeringDB facility coordinates now get geocoded via:
1. City name extraction from IX name (e.g. "France-IX Paris" → Paris)
2. Hard-coded IX ID → city map for 15 well-known IXPs (SwissIX→Zurich etc.)
3. 70+ major networking cities with lat/lon coordinates
AS8283 Coloclue: 9 → 12 IX locations (5 cities: AMS, FRA, Paris, Zurich, Meppel)
AS49544 i3D.net: 100 connections → 20 locations (16 cities worldwide)
- Leaflet map: double requestAnimationFrame after display:none removal
ensures container has real dimensions before L.map() init
- PeeringDB org cache: 24h disk cache (.pdb-org-cache.json) prevents
hammering PeeringDB API on server restarts (was causing 175 restarts)
- Check HTTP status before JSON.parse on PDB responses
- Leaflet.js (CDN) with CartoDB Dark Matter tiles matching Tokyo Night theme
- Cyan markers: facility/datacenter locations with name + city popup
- Orange markers: IX presence with IX name + speed popup
- Purple connecting lines between facilities in the same country
- Coordinates from PeeringDB facility API (batch lookup, chunked)
- IX locations via ixfac association + facility geocoding
- Auto-fit bounds, graceful degradation if no coordinates
- Collapsible card, XSS-safe popups via DOM API
- RPKI cross-check: Cloudflare RPKI feed + RIPE NCC Validator API (5 sample prefixes)
- Prefix cross-check: RIPE Stat vs bgp.he.net count comparison
- Neighbour cross-check: RIPE Stat vs bgp.he.net peer data
- Data Quality badge in dashboard (High/Medium/Low confidence)
- Hover tooltip: "Data Quality Report" with per-source agreement breakdown
- Added BETA tag to site header and version string (v0.5.0-beta)
- All UI text in English
RPKI Validation:
- Validate ALL prefixes (not sample of 10) using local Cloudflare RPKI feed
- Covers all 5 RIRs globally (RIPE, APNIC, ARIN, LACNIC, AFRINIC)
- Indexed ROA lookup (O(bucket) not O(824K)) for instant validation
- AS4739 now correctly shows 446/446 prefixes checked
ASPA Provider Detection:
- Only RIPE Stat "left" neighbours (verified upstreams) used as providers
- AS-path analysis used for frequency confirmation only, not as provider source
- Fixes false provider detection that included peers alongside upstreams
Multi-RIR Support:
- WHOIS/IRR queries all 5 RIR databases via RDAP in parallel
- RPSL validation checks RIPE + APNIC/ARIN/LACNIC/AFRINIC
- AS4739 (APNIC) now correctly found via rdap.apnic.net
Geolocation:
- Anycast/CDN networks (5+ facility countries or Content/NSP type) not flagged
- Only small networks with geo anomalies get warnings
Route Server Scoring:
- Networks with 20+ IX connections and no RS scored as "pass" (bilateral policy)
- Only small networks without RS get warnings
Error Handling:
- ASPA endpoints gracefully handle timeouts (show fallback instead of HTML parse error)
- Frontend checks Content-Type before JSON.parse
Reported by Philip Smith, Richard Steenbergen, Jared Mauch, Chris Malayter
- Bug 1: Facilities returned ALL 59k PeeringDB entries (missing net_id filter)
- Bug 2: Neighbours returned 0 for large ASNs (8s timeout, now 30s)
- Bug 3: Visibility showed 0% when API times out (now shows -1 = unavailable)
- Bug 4: Prefixes returned 0 for small ASNs (cascading fetch failure)
- Bug 5: RPKI inconsistency documented (different sample sets per endpoint)
- Bug 6: Atlas probe status showed 0 connected (status.name vs status_name)
Also: PDB IX/Fac queries now use net_id (phase 0 + parallel phase 1)
Also: Compare endpoint uses net_id for facilities
- Replace static screenshot images with links to live demo at peercortex.org
- Add "What's New in v0.5.0" section with all new features
- Add ASPA Verification section (draft-ietf-sidrops-aspa-verification compliance)
- Add Cloudflare RPKI JSON feed and Route Views to Data Sources table
- Update tool count from 25+ to 34 MCP tools, 12 REST API endpoints
- Bump version from 0.1.0 to 0.5.0 in package.json
- Compare: all API calls in single parallel batch (was sequential)
- Compare: RPKI sample reduced to 3+3 prefixes with 5s timeout cap
- Compare: response caching (5min TTL)
- Compare: AS name resolution parallel with 3s timeout
- Result: Compare from timeout (>20s) to ~5s first call, <1s cached
- Add response caching (5min TTL for lookups, 10min for ASPA)
- Add 8s timeout to all external API fetches
- RPKI validation: sample max 10 prefixes (5 v4 + 5 v6) instead of 50
- Run all PeeringDB + RIPE Stat calls in single parallel batch
- Resolve AS names in parallel with 3s timeout cap
- Add deploy/ directory with production server.js + index.html dashboard
- Landing page: Tokyo Night dark theme, interactive ASN search
- 15 API endpoints: lookup, aspa, aspa/verify, bgproutes, validate,
compare, peers/find, prefix/detail, ix/detail, topology, whois, health
- Features: RPKI per-prefix, RIPE Atlas probes, Network Health Report,
ASPA RFC verification engine, Provider Relationship Graph
- Query RIPE Atlas API for probes in the looked-up ASN
- Display probe count, connected/disconnected status, anchors
- Expandable probe detail table with links to atlas.ripe.net
- Connection ratio progress bar
- "Host a probe?" prompt for networks without Atlas presence
PeerCortex unifies PeeringDB, RIPE Stat, bgproutes.io, RIPE Atlas,
Route Views, IRR, RPKI, and CAIDA into a single AI-queryable MCP Server
for network engineers. Powered by local Ollama.
Core capabilities:
- 34 MCP tools for network intelligence
- 11 data sources unified
- ASPA validation engine (RFC 9582) with leak detection
- Peering partner discovery with AI-ranked matches
- BGP analysis and anomaly detection
- RPKI monitoring and compliance reports
- Latency/traceroute via RIPE Atlas
- Transit analysis and cost comparison
- IX traffic statistics
- AS topology mapping
- ASPA object generator and simulator
- 100% local AI — no cloud dependencies
