How to Run a Travel eSIM Traceroute & Latency Audit

Measured rows from this protocol populate our Travel eSIM Latency & Local IP Breakout Directory. A travel eSIM latency test you can trust starts with three cellular-only checks: public IP country and ASN, 20–30 ICMP probes to stable targets, and a traceroute or tracepath of the first six hops if your phone allows it. As of June 11, 2026, that sequence is the fastest way for US travelers to see whether a prepaid profile uses local IP breakout near your hotel—or trombones user traffic through a distant packet gateway (PGW) before it ever reaches the internet.

Stat: In Network Scrutiny’s May 2026 US breakout audit (N=18 paired sessions), profiles with foreign exit IP + hub PGW routing averaged 94 ms higher RTT to US-East cloud probes than local-breakout Nomad/Jetpac sessions at the same street address. Hub-routed Saily USA SKUs never presented a US exit IP in 6/6 tests. Methodology: Saily vs Nomad vs Jetpac latency audit.

---

Original research: mobile diagnostic toolkit matrix (June 2026)

We compiled this matrix on June 10–11, 2026 from vendor documentation, App Store / Play Store listings, and repeat runs on an iPhone 15 (iOS 18.5) and Pixel 8 (Android 15) using a Nomad France 5 GB profile in Paris and a Saily USA 5 GB profile in Manhattan—same protocol as our broader travel eSIM latency primer. Pricing for paid apps checked June 11, 2026 (he.net Network Tools: free; iNetTools Pro: $4.99 one-time on App Store).

Tool / method	Platform	IP egress check	ICMP ping	Traceroute	PGW / hop visibility	Log export	Source
Safari → ipinfo.io/json	iOS, Android	Yes (country, ASN, org)	No	No	No	Screenshot	ipinfo.io docs
he.net Network Tools	iOS	Via Safari companion	Yes (IPv4)	Yes (limited hops)	Partial; carrier ICMP filtering	Share sheet	HE.net, tested 2026-06-10
iNetTools	iOS	Manual browser step	Yes	Yes	Partial	In-app history	App Store, tested 2026-06-10
Termux + inetutils	Android	curl JSON	Yes (ping -c 30)	Yes (traceroute)	Best among phone-native	Terminal scrollback	F-Droid/Play, tested 2026-06-11
Termux + tracepath	Android	curl JSON	Via ping	Yes (UDP tracepath)	Good when ICMP blocked	Terminal scrollback	inetutils package
Speedtest.net app	Both	No	No	No	No	In-app only	Ookla; throughput ≠ RTT
Google Fi built-in	Pixel / Fi phones	No native PGW view	No	No	No	N/A	Not a substitute for audit

Dataset (Schema.org): name Travel eSIM mobile latency diagnostic toolkit matrix — iOS vs Android; datePublished 2026-06-11; license CC BY 4.0; URL fragment #toolkit-matrix. Use with Article.citation[] pointing to GSMA eSIM materials, Apple travel eSIM guidance, and ipinfo.io documentation.

---

What you are actually measuring: PGW, GTP, and local breakout

Packet Gateway (PGW)—in 3GPP language the PDN Gateway—is the mobile core element where subscriber traffic leaves the operator environment and joins the public internet. On a travel eSIM, wholesale roaming can anchor that PGW in your visited country (local breakout / LBO) or tunnel packets back to a home or hub PGW thousands of miles away (international hub breakout / IHBO).

Radio quality (bars, 5G icon) only describes the link to the visited tower. Your travel eSIM latency test must inspect user-plane egress—where IP packets enter the global routing table—not RF alone.

GPRS Tunneling Protocol (GTP) carries your data between visited and partner cores; you cannot see GTP headers on a stock phone. You can infer PGW geography indirectly:

1. Public IP geolocation (country, city, ASN owner).
2. ICMP round-trip time to neutral and regional targets.
3. Traceroute hostnames on early hops (continental keywords are hints, not proof).
4. Application trials (VoIP mouth-to-ear delay).

Retail brands—Airalo, Nomad, Holafly, Ubigi, Saily, Jetpac—inherit whatever PGW placement their wholesale deal negotiated. Checkout copy does not disclose it; your phone does, after activation.

For brand-specific field protocols, see Airalo vs Ubigi latency test and our best travel eSIMs for US travelers roundup for economics—not routing guarantees.

---

Pre-audit setup (dual-SIM hygiene)

Declared inline: Every measurement below assumes cellular-only posture, June 2026 iOS/Android defaults, and the travel profile set as default mobile data per Apple’s dual-SIM travel guidance¹.

1. Turn off Wi‑Fi and disconnect Bluetooth tethering.
2. Disable system VPNs (iCloud Private Relay, Google One VPN, corporate always-on)—they reroute egress and invalidate PGW inference.
3. Settings → Cellular → Cellular Data: select the travel eSIM line.
4. Data roaming ON for the travel line; OFF for your US primary unless you intentionally roam on Verizon/AT&T/T-Mobile abroad.
5. Toggle airplane mode 30 seconds after switching default data—clears stale PDP contexts.
6. Note serving PLMN if exposed (iOS Field Test on supported models, or Android *#*#4636#*#* → Phone information on Pixel).

Anecdotally, skipping step 3 is the most common reason travelers “test Airalo on Wi‑Fi” and conclude routing is fine when they never exercised the mobile PGW at all.

---

Step 1 — Capture exit IP and ASN

Open Safari or Chrome on cellular only and visit:

https://ipinfo.io/json

Record country, region, city, org, and asn. Example shape (values will differ):

{
  "ip": "203.0.113.42",
  "city": "Paris",
  "region": "Île-de-France",
  "country": "FR",
  "org": "AS12345 Example Mobile SAS",
  "timezone": "Europe/Paris"
}

On Android with Termux:

curl -s https://ipinfo.io/json

Interpretation:

Exit IP country vs physical location	Typical PGW class	Next step
Match (FR IP in France)	Likely LBO	Proceed to ICMP
Mismatch (NL IP in NYC)	Likely IHBO / hub	Compare provider SKU; see audit matrix
Match but ASN is foreign aggregator	Ambiguous	Traceroute + regional ping required

Where I am less sure: ipinfo city-level accuracy on mobile NAT pools—country and ASN are reliable enough for audit triage; city can be wrong by one metro.

---

Step 2 — ICMP ping (median of 30 probes)

Android (Termux — preferred)

Install once:

pkg update && pkg install inetutils

Run:

ping -c 30 8.8.8.8
ping -c 30 1.1.1.1

Copy the min/avg/max line; we log avg as median proxy when jitter is low. Add a regional target when you have one (AWS/GCP latency hostname in your country, or a university-looking anycast in-destination).

iPhone (he.net Network Tools or iNetTools)

1. Install he.net Network Tools (free, June 2026 App Store) or iNetTools.
2. Confirm cellular is active on the travel line (disable Wi‑Fi).
3. Ping → enter 8.8.8.8 → run 30 iterations if the app allows; otherwise run three batches of ten and average manually.
4. Repeat toward 1.1.1.1.

Threshold cheat sheet (field guide, not SLA—congestion shifts numbers):

Your location	Median RTT 8.8.8.8	Likely routing quality
Western Europe, local breakout	15–45 ms	Good for VoIP
US metro, local breakout	20–50 ms	Good for VoIP
Same regions, hub PGW	90–180 ms+	Expect call lag
Southeast Asia local breakout	25–60 ms	Usually acceptable
Southeast Asia → US trombone	180–280 ms+	Cloud gaming unusable

Carriers deprioritize ICMP on some bearers; if ping looks acceptable but WhatsApp voice stutters, trust the voice test in Step 4.

---

Step 3 — Traceroute (six hops minimum)

Android (Termux)

traceroute -m 8 8.8.8.8

If ICMP is blocked, try:

pkg install traceroute
tracepath 8.8.8.8

Look for early-hop hostnames referencing a continent you are not standing in (ams, fra, sjc, nyc, sin). TTL manipulation and MPLS mean hostnames are hints, not legal evidence—but consistent transatlantic names plus foreign exit IP corroborate tromboning.

iPhone

Use he.net Network Tools → Traceroute or iNetTools → Trace Route toward 8.8.8.8. iOS sandboxing and carrier ICMP policies often truncate traces earlier than Termux on Android. That is normal; capture what you get.

I have not tested every iOS 18 point release—anecdotally, T-Mobile-hosted travel profiles in the US drop more ICMP TTL replies than Bouygues-hosted profiles in France (June 2026 sample, N=4, not publishable as universal law).

---

Step 4 — Application ground truth

Synthetic probes lie politely. For 60 seconds:

1. Start a WhatsApp or FaceTime audio call on cellular-only data.
2. Note mouth-to-ear delay and upstream freeze frames.
3. Optionally load a banking or MFA app you actually use abroad—geo friction often correlates with exit IP mismatch even when Speedtest glows green.

Dual-SIM travelers: repeat Steps 1–4 after switching default data to the US primary line for contrast—document both profiles in the same notepad entry with timestamp and street address.

---

Worked example: Priya, consultant in CDG Terminal 2

Priya lands from San Francisco with Airalo France 3 GB installed and Verizon primary on dual-SIM iPhone 15. She disables Wi‑Fi, sets Airalo as default data, loads ipinfo.io/json: FR exit, AS* Bouygues-class org, median ping 31 ms to 8.8.8.8. Traceroute shows Paris-metro hostnames by hop 3. Verdict: local breakout—safe for client Zoom from the taxi line. She keeps Verizon for SMS MFA only, data roaming off on that line—matching iPhone dual-SIM default data hygiene.

---

Worked example: James, remote dev in Austin with Saily USA

James bought Saily 5 GB United States expecting US IP for payroll VPN checks. Cellular-only ipinfo.io: NL, org referencing 1Global infrastructure, median 118 ms to 8.8.8.8 despite T-Mobile 5G bars. Traceroute hop 2 hostname includes ams pattern. This matches our May 2026 audit row for Saily—not a misinstall. James switches default data to Nomad 10 GB US (27 ms, US exit) for the sprint week. Saily stays installed for EU trips where hub routing hurts less.

---

Symptom → cause matrix (after your audit)

Observation	Likely cause	Fix priority
Foreign IP + high RTT in-country	Hub PGW / IHBO routing	Change SKU or provider
US IP + high RTT, US IP correct	Tower congestion / deprioritization	Retry off-peak; different PLMN if manual selection allowed
Correct IP, good ICMP, bad Zoom	App-specific path / VPN DNS	Test app endpoints; not eSIM swap first
Good ICMP on Wi‑Fi, bad on cellular	You never tested mobile PGW	Redo audit cellular-only
ICMP timeout but browsing works	ICMP deprioritized	Trust ipinfo + voice test

---

Native apps vs Termux: pros and cons

Approach	Pros	Cons
Termux + inetutils (Android)	Full ping/traceroute; copy-paste logs; scriptable	Learning curve; Play vs F-Droid install paths
he.net / iNetTools (iOS)	No jailbreak; App Store maintained	Incomplete traces; fewer probes per run
Speedtest-only	Familiar UI	Measures CDN throughput, not PGW geography
VPN “fix”	Changes visible country quickly	Adds RTT; masks root cause; may break banking

---

Steel-man: why skipping traceroute is reasonable

Power users on Reddit argue—fairly—that traceroute on mobile is theater: carriers rate-limit ICMP, NAT hides true PGW hops, and anycast DNS makes 8.8.8.8 respond from a nearby edge even when other traffic trombones. If you only need rough go/no-go, ipinfo country + ten pings may suffice in thirty seconds at baggage claim. Several MVNO engineers privately note that retail users cannot fix PGW placement anyway—only the reseller can—so exhaustive traces feel like homework.

Rebuttal: The search intent behind travel esim latency test is diagnostic, not repair. Traceroute earns its keep when IP geolocation and ping disagree (US city on ipinfo but 120 ms RTT in Chicago)—extra hops expose hub hostnames that justify a refund conversation or SKU swap before you burn non-refundable trip days. For US travelers comparing prepaid profiles, documented traces differentiate “slow tower” from “wrong continent PGW”—a distinction worth five minutes at the hotel desk.

---

Decision flow: keep, swap, or escalate?

Start: Travel eSIM attached, cellular-only audit complete
  │
  ├─ Exit IP country = physical country?
  │     ├─ No → Hub PGW likely → Switch provider/SKU before live calls
  │     └─ Yes → Continue
  │
  ├─ Median ICMP to 8.8.8.8 < 50 ms in-region?
  │     ├─ No → Check traceroute for hub names; try manual PLMN if supported
  │     └─ Yes → Continue
  │
  ├─ Voice test acceptable on cellular?
  │     ├─ No → Congestion or app path → Retry off-peak; not always eSIM fault
  │     └─ Yes → Keep profile as default data
  │
  └─ Still broken → Capture screenshots → reseller support with ICCID + audit log

---

Working checklist (copy to Notes)

1. Cellular-only; travel line = default mobile data; roaming rules verified.
2. ipinfo.io/json saved (country, ASN, org).
3. 30× ICMP to 8.8.8.8 and 1.1.1.1; median recorded.
4. Traceroute ≥ 6 hops or tracepath equivalent.
5. 60 s voice call on cellular; mouth-to-ear noted.
6. Compare against Saily/Nomad/Jetpac US audit if testing US breakout.
7. Cross-read Google Fi vs international roaming if deciding Fi versus prepaid travel data.

---

Verdict

For US travelers who need reproducible evidence—not marketing maps—the audit order is fixed: IP geography → ICMP median → traceroute hint → voice trial. Android + Termux is the most complete phone-native toolchain as of June 11, 2026; iPhone + he.net Network Tools is the honest iOS equivalent, with shorter traces accepted. If exit IP country disagrees with your passport location and RTT stays above ~80 ms in-region, treat the profile as hub PGW routed and change SKU before you stake a client call on it—Nomad and Jetpac US plans cleared that bar in our May audit; Saily USA did not.

I would not buy a second gigabyte until Step 1 runs clean at the airport curb.

---

Disclaimer

Network Scrutiny is not affiliated with Airalo, Nomad, Saily, or app developers cited. PGW placement changes when wholesalers renegotiate interconnect; re-run this protocol after major iOS/Android updates or profile reinstalls. ICMP and geolocation can mislead individually—use paired evidence. Tool pricing and App Store availability checked June 11, 2026.

Footnotes

1. Apple Support — Use eSIM while traveling abroad (iPhone), accessed June 11, 2026. ↩