Lower RTT from Warsaw and Milan
All three probes in each selection favoured NDC2. City medians were 21.98 vs 32.59 ms from Warsaw, and 15.76 vs 24.17 ms from Milan.
FIELD REPORT 01 · 17 SEPTEMBER 2026
Vienna NDC2 vs Interxion.
Let the routes speak.
A paired look at two Austrian VPS locations from 16 European source selections. Every result, including the inconvenient ones, is available to inspect.
Historical VPS sample: the temporary test instances have been retired. For new tests, use the permanent Looking Glass addresses. They were not the targets of this report.
47 / 48requested probe pairs returned
16ping packets per target / probe
752 / 752NDC2 ping replies
748 / 752Interxion ping replies
What this first sample tells us
Two useful options.
No universal winner.
This is one collection lasting about 12 minutes, not continuous monitoring. Treat the differences as leads for testing your own audience.
Lower RTT from Vienna and Bratislava
All three probes in each selection favoured Interxion. City medians were 0.66 vs 1.27 ms from Vienna, and 2.04 vs 13.05 ms from Bratislava.
One city can hide very different paths
Frankfurt’s medians differ by just 0.04 ms. Its OVH probe, however, favoured NDC2 by 18.53 ms. Budapest’s ATW probe also favoured NDC2 despite Interxion’s lower city median.
Packet loss is included. NDC2 returned all 752 ping packets; Interxion missed 4 of 752 (0.53%): 2 from one Bratislava probe, 1 from Budapest and 1 from Munich. Separately, MTR destination replies missed 1/752 at NDC2 and 6/752 at Interxion. These short samples cannot establish a sustained loss rate or its cause.
The complete comparison
Latency into Vienna.
Each city value is the median of the individual probes’ average ping RTTs. It is not a city-wide average or the median of all packets. Ljubljana has two probes; the other selections have three.
Vienna NDC2Vienna Interxion
Median of per-probe averages, milliseconds. All bars share a 50 ms scale.
View exact city values and ranges
| Source selection | Probes | NDC2 | Interxion | Range of probe averages NDC2 / Interxion |
|---|---|---|---|---|
| Amsterdam | 3 | 22.32 | 18.12 | 17.44–24.66 / 17.91–22.45 |
| Bratislava | 3 | 13.05 | 2.04 | 3.38–13.23 / 1.99–2.48 |
| Budapest | 3 | 14.83 | 5.83 | 6.21–24.55 / 5.34–25.89 |
| Frankfurt | 3 | 11.83 | 11.79 | 11.50–13.01 / 11.52–31.54 |
| Ljubljana | 2 | 7.42 | 7.58 | 7.36–7.48 / 7.50–7.67 |
| London | 3 | 28.50 | 23.55 | 22.71–28.62 / 22.56–24.32 |
| Milan | 3 | 15.76 | 24.16 | 11.79–47.71 / 24.06–56.88 |
| Munich | 3 | 9.46 | 8.85 | 1.49–26.44 / 1.20–33.15 |
| Paris | 3 | 23.67 | 23.19 | 20.55–27.39 / 21.25–33.49 |
| Prague | 3 | 11.37 | 7.85 | 8.07–17.52 / 6.89–18.97 |
| Stockholm | 3 | 31.24 | 34.30 | 29.72–134.00 / 31.56–142.66 |
| Tallinn | 3 | 41.70 | 44.75 | 39.11–42.13 / 26.85–51.89 |
| Vienna | 3 | 1.27 | 0.66 | 1.07–8.38 / 0.63–7.13 |
| Warsaw | 3 | 21.98 | 32.59 | 21.95–24.64 / 32.56–40.35 |
| Zagreb | 3 | 10.01 | 9.13 | 9.53–30.22 / 8.72–29.59 |
| Zurich | 3 | 17.42 | 17.51 | 15.65–33.91 / 16.17–22.78 |
Source city labels are reported by Globalping, not independently geolocated by EDIS Global. The full ranges and individual measurements remain visible, including the high-variation Stockholm probe.
Behind the numbers
Different paths
into the same city.
MTR reveals visible networks along the forward path. Round-trip latency also depends on the return path, which this collection does not reveal.
NDC2: Nessus to EDIS Global
The sampled NDC2 traces commonly pass through Nessus AS47692 before the target attributed to EDIS Global AS57169. The exact upstream sequence changes with the source network.
Interxion: M247 network path
The sampled Interxion traces commonly reach AS9009, associated with M247. This differs from NDC2’s observed network path; it does not establish the sites’ complete upstream contracts or failure independence.
Warsaw · OVH AS16276
NDC2
- AS16276
- AS47692
- AS57169
Interxion
- AS16276
- AS9009
Consecutive repeated AS numbers collapsed; non-responding or unattributed hops omitted. This is the visible MTR sequence, not a complete BGP path. NDC2 raw MTR · Interxion raw MTR
The Warsaw OVH probe reaches NDC2 through the visible Nessus/EDIS sequence, while the Interxion path is attributed to M247. The accompanying ping means were 24.64 ms and 40.35 ms respectively. Fewer visible AS numbers did not mean lower latency.
Budapest · M247 AS9009
NDC2
- AS9009
- AS201011
- AS47692
- AS57169
Interxion
- AS9009
Consecutive repeated AS numbers collapsed; non-responding or unattributed hops omitted. This is the visible MTR sequence, not a complete BGP path. NDC2 raw MTR · Interxion raw MTR
From the M247 probe in Budapest, Interxion’s visible path stays within AS9009 and its ping mean was 5.34 ms, compared with 24.55 ms to NDC2. This is an on-network example, not evidence that every Hungarian network favours Interxion. The ATW probe showed the opposite result: 14.83 ms to NDC2 and 25.89 ms to Interxion.
Routers may delay or omit ICMP replies while forwarding traffic normally. Intermediate-hop loss and hostnames alone do not prove congestion, a physical detour, a peering agreement or a fault. Use the destination statistics and repeat the test.
Check the evidence
Every probe.
Both destinations.
Expand a source selection to see the network, average ping RTT, missing replies and original outputs. No probe was removed for being slow or showing loss.
Amsterdam3 paired probes · ping & MTR
Reported source networks and AS numbers. Ping shows the mean of 16 attempts; loss is the number of missing replies. MTR values below are destination-only.
| Source network | NDC2 ping | Interxion ping | NDC2 / Interxion MTR destination loss |
|---|---|---|---|
| OracleAmsterdam, NL · AS31898 | 22.32 ms0/16 lost (0%) | 17.91 ms0/16 lost (0%) | 0/16 / 0/16 |
| LeaseWebAmsterdam, NL · AS60781 | 17.44 ms0/16 lost (0%) | 18.12 ms0/16 lost (0%) | 0/16 / 0/16 |
| DELUXHOSTAmsterdam, NL · AS214677 | 24.66 ms0/16 lost (0%) | 22.45 ms0/16 lost (0%) | 0/16 / 0/16 |
Bratislava3 paired probes · ping & MTR
Reported source networks and AS numbers. Ping shows the mean of 16 attempts; loss is the number of missing replies. MTR values below are destination-only.
| Source network | NDC2 ping | Interxion ping | NDC2 / Interxion MTR destination loss |
|---|---|---|---|
| CastlegemBratislava, SK · AS34689 | 13.05 ms0/16 lost (0%) | 1.99 ms0/16 lost (0%) | 0/16 / 0/16 |
| PE Denis PodolskiiBratislava, SK · AS61335 | 3.38 ms0/16 lost (0%) | 2.48 ms2/16 lost (12.5%) | 1/16 / 1/16 |
| eServerBratislava, SK · AS61424 | 13.23 ms0/16 lost (0%) | 2.04 ms0/16 lost (0%) | 0/16 / 2/16 |
Budapest3 paired probes · ping & MTR
Reported source networks and AS numbers. Ping shows the mean of 16 attempts; loss is the number of missing replies. MTR values below are destination-only.
| Source network | NDC2 ping | Interxion ping | NDC2 / Interxion MTR destination loss |
|---|---|---|---|
| Magyar TelekomBudapest, HU · AS5483 | 6.21 ms0/16 lost (0%) | 5.83 ms0/16 lost (0%) | 0/16 / 0/16 |
| M247Budapest, HU · AS9009 | 24.55 ms0/16 lost (0%) | 5.34 ms1/16 lost (6.25%) | 0/16 / 0/16 |
| ATW InternetBudapest, HU · AS41075 | 14.83 ms0/16 lost (0%) | 25.89 ms0/16 lost (0%) | 0/16 / 0/16 |
Frankfurt3 paired probes · ping & MTR
Reported source networks and AS numbers. Ping shows the mean of 16 attempts; loss is the number of missing replies. MTR values below are destination-only.
| Source network | NDC2 ping | Interxion ping | NDC2 / Interxion MTR destination loss |
|---|---|---|---|
| OVHFrankfurt, DE · AS16276 | 13.01 ms0/16 lost (0%) | 31.54 ms0/16 lost (0%) | 0/16 / 0/16 |
| OracleFrankfurt, DE · AS31898 | 11.83 ms0/16 lost (0%) | 11.79 ms0/16 lost (0%) | 0/16 / 0/16 |
| synlinq.deFrankfurt, DE · AS44486 | 11.50 ms0/16 lost (0%) | 11.52 ms0/16 lost (0%) | 0/16 / 0/16 |
Ljubljana2 paired probes · ping & MTR
Reported source networks and AS numbers. Ping shows the mean of 16 attempts; loss is the number of missing replies. MTR values below are destination-only.
| Source network | NDC2 ping | Interxion ping | NDC2 / Interxion MTR destination loss |
|---|---|---|---|
| Optimus ITLjubljana, SI · AS48894 | 7.48 ms0/16 lost (0%) | 7.67 ms0/16 lost (0%) | 0/16 / 1/16 |
| HostRoyale TechnologiesLjubljana, SI · AS203020 | 7.36 ms0/16 lost (0%) | 7.50 ms0/16 lost (0%) | 0/16 / 1/16 |
London3 paired probes · ping & MTR
Reported source networks and AS numbers. Ping shows the mean of 16 attempts; loss is the number of missing replies. MTR values below are destination-only.
| Source network | NDC2 ping | Interxion ping | NDC2 / Interxion MTR destination loss |
|---|---|---|---|
| DigitalOceanLondon, GB · AS14061 | 28.50 ms0/16 lost (0%) | 23.55 ms0/16 lost (0%) | 0/16 / 0/16 |
| OVHLondon, GB · AS16276 | 28.62 ms0/16 lost (0%) | 22.56 ms0/16 lost (0%) | 0/16 / 0/16 |
| OracleLondon, GB · AS31898 | 22.71 ms0/16 lost (0%) | 24.32 ms0/16 lost (0%) | 0/16 / 0/16 |
Milan3 paired probes · ping & MTR
Reported source networks and AS numbers. Ping shows the mean of 16 attempts; loss is the number of missing replies. MTR values below are destination-only.
| Source network | NDC2 ping | Interxion ping | NDC2 / Interxion MTR destination loss |
|---|---|---|---|
| LAKENETWORKSMilan, IT · AS6517 | 15.76 ms0/16 lost (0%) | 24.16 ms0/16 lost (0%) | 0/16 / 0/16 |
| FastwebMilan, IT · AS12874 | 47.71 ms0/16 lost (0%) | 56.88 ms0/16 lost (0%) | 0/16 / 1/16 |
| Akamai Connected CloudMilan, IT · AS63949 | 11.79 ms0/16 lost (0%) | 24.06 ms0/16 lost (0%) | 0/16 / 0/16 |
Munich3 paired probes · ping & MTR
Reported source networks and AS numbers. Ping shows the mean of 16 attempts; loss is the number of missing replies. MTR values below are destination-only.
| Source network | NDC2 ping | Interxion ping | NDC2 / Interxion MTR destination loss |
|---|---|---|---|
| Deutsche TelekomMunich, DE · AS3320 | 9.46 ms0/16 lost (0%) | 8.85 ms1/16 lost (6.25%) | 0/16 / 0/16 |
| Leibniz-RechenzentrumMunich, DE · AS12816 | 26.44 ms0/16 lost (0%) | 33.15 ms0/16 lost (0%) | 0/16 / 0/16 |
| Freie Netze MuenchenMunich, DE · AS212567 | 1.49 ms0/16 lost (0%) | 1.20 ms0/16 lost (0%) | 0/16 / 0/16 |
Paris3 paired probes · ping & MTR
Reported source networks and AS numbers. Ping shows the mean of 16 attempts; loss is the number of missing replies. MTR values below are destination-only.
| Source network | NDC2 ping | Interxion ping | NDC2 / Interxion MTR destination loss |
|---|---|---|---|
| SCALEWAYParis, FR · AS12876 | 27.39 ms0/16 lost (0%) | 33.49 ms0/16 lost (0%) | 0/16 / 0/16 |
| The Constant CompanyParis, FR · AS20473 | 20.55 ms0/16 lost (0%) | 21.25 ms0/16 lost (0%) | 0/16 / 0/16 |
| Layer7 NetworksParis, FR · AS199912 | 23.67 ms0/16 lost (0%) | 23.19 ms0/16 lost (0%) | 0/16 / 0/16 |
Prague3 paired probes · ping & MTR
Reported source networks and AS numbers. Ping shows the mean of 16 attempts; loss is the number of missing replies. MTR values below are destination-only.
| Source network | NDC2 ping | Interxion ping | NDC2 / Interxion MTR destination loss |
|---|---|---|---|
| Cogent CommunicationsPrague, CZ · AS174 | 17.52 ms0/16 lost (0%) | 6.89 ms0/16 lost (0%) | 0/16 / 0/16 |
| T-MobilePrague, CZ · AS13036 | 8.07 ms0/16 lost (0%) | 7.85 ms0/16 lost (0%) | 0/16 / 0/16 |
| INTERNET CZPrague, CZ · AS24806 | 11.37 ms0/16 lost (0%) | 18.97 ms0/16 lost (0%) | 0/16 / 0/16 |
Stockholm3 paired probes · ping & MTR
Reported source networks and AS numbers. Ping shows the mean of 16 attempts; loss is the number of missing replies. MTR values below are destination-only.
| Source network | NDC2 ping | Interxion ping | NDC2 / Interxion MTR destination loss |
|---|---|---|---|
| IncogNETStockholm, SE · AS40663 | 29.72 ms0/16 lost (0%) | 34.30 ms0/16 lost (0%) | 0/16 / 0/16 |
| HostHatchStockholm, SE · AS63473 | 31.24 ms0/16 lost (0%) | 31.56 ms0/16 lost (0%) | 0/16 / 0/16 |
| Hollander & JacobsenStockholm, SE · AS215960 | 134.00 ms0/16 lost (0%) | 142.66 ms0/16 lost (0%) | 0/16 / 0/16 |
Tallinn3 paired probes · ping & MTR
Reported source networks and AS numbers. Ping shows the mean of 16 attempts; loss is the number of missing replies. MTR values below are destination-only.
| Source network | NDC2 ping | Interxion ping | NDC2 / Interxion MTR destination loss |
|---|---|---|---|
| xTomTallinn, EE · AS3214 | 39.11 ms0/16 lost (0%) | 51.89 ms0/16 lost (0%) | 0/16 / 0/16 |
| Baykov Ilya SergeevichTallinn, EE · AS41745 | 41.70 ms0/16 lost (0%) | 44.75 ms0/16 lost (0%) | 0/16 / 0/16 |
| TAKEHOSTTallinn, EE · AS214639 | 42.13 ms0/16 lost (0%) | 26.85 ms0/16 lost (0%) | 0/16 / 0/16 |
Vienna3 paired probes · ping & MTR
Reported source networks and AS numbers. Ping shows the mean of 16 attempts; loss is the number of missing replies. MTR values below are destination-only.
| Source network | NDC2 ping | Interxion ping | NDC2 / Interxion MTR destination loss |
|---|---|---|---|
| T-MobileVienna, AT · AS8412 | 8.38 ms0/16 lost (0%) | 7.13 ms0/16 lost (0%) | 0/16 / 0/16 |
| Hohl ITVienna, AT · AS40994 | 1.07 ms0/16 lost (0%) | 0.63 ms0/16 lost (0%) | 0/16 / 0/16 |
| netcupVienna, AT · AS197540 | 1.27 ms0/16 lost (0%) | 0.66 ms0/16 lost (0%) | 0/16 / 0/16 |
Warsaw3 paired probes · ping & MTR
Reported source networks and AS numbers. Ping shows the mean of 16 attempts; loss is the number of missing replies. MTR values below are destination-only.
| Source network | NDC2 ping | Interxion ping | NDC2 / Interxion MTR destination loss |
|---|---|---|---|
| OVHWarsaw, PL · AS16276 | 24.64 ms0/16 lost (0%) | 40.35 ms0/16 lost (0%) | 0/16 / 0/16 |
| AlbHostWarsaw, PL · AS48014 | 21.98 ms0/16 lost (0%) | 32.59 ms0/16 lost (0%) | 0/16 / 0/16 |
| MEVSPACEWarsaw, PL · AS201814 | 21.95 ms0/16 lost (0%) | 32.56 ms0/16 lost (0%) | 0/16 / 0/16 |
Zagreb3 paired probes · ping & MTR
Reported source networks and AS numbers. Ping shows the mean of 16 attempts; loss is the number of missing replies. MTR values below are destination-only.
| Source network | NDC2 ping | Interxion ping | NDC2 / Interxion MTR destination loss |
|---|---|---|---|
| A1Zagreb, HR · AS12810 | 30.22 ms0/16 lost (0%) | 29.59 ms0/16 lost (0%) | 0/16 / 0/16 |
| A1Zagreb, HR · AS31012 | 9.53 ms0/16 lost (0%) | 9.13 ms0/16 lost (0%) | 0/16 / 0/16 |
| DatacampZagreb, HR · AS60068 | 10.01 ms0/16 lost (0%) | 8.72 ms0/16 lost (0%) | 0/16 / 0/16 |
Zurich3 paired probes · ping & MTR
Reported source networks and AS numbers. Ping shows the mean of 16 attempts; loss is the number of missing replies. MTR values below are destination-only.
| Source network | NDC2 ping | Interxion ping | NDC2 / Interxion MTR destination loss |
|---|---|---|---|
| SunriseZurich, CH · AS6730 | 15.65 ms0/16 lost (0%) | 22.78 ms0/16 lost (0%) | 0/16 / 0/16 |
| OracleZurich, CH · AS31898 | 17.42 ms0/16 lost (0%) | 17.51 ms0/16 lost (0%) | 0/16 / 0/16 |
| HostHatchZurich, CH · AS63473 | 33.91 ms0/16 lost (0%) | 16.17 ms0/16 lost (0%) | 0/16 / 0/16 |
Methodology & limitations
A documented snapshot.
Measurements were collected by EDIS Global using Globalping’s distributed probes. EDIS Global publishes this report and sells the services compared.
EDIS Global also supports Globalping with infrastructure. The results are an operator-published sample, not an independent certification or endorsement by Globalping.
Targets and source selection
The targets were temporary VPS instances at NDC2 (AT) and Interxion (AT-VIE), which have since been retired. Their original addresses remain in the archived evidence for traceability and must not be used for new tests. The results do not describe the current Looking Glass endpoints.
The initial ping selected up to three probes per source. Its measurement ID was reused for the second target and both MTR tests. Returned probe metadata matches across the four measurements in each selection.
Time, order and coverage
17 September 2026, 10:17:00–10:29:25 UTC (12:17–12:29 CEST). In this first run, AT preceded AT-VIE; ping preceded MTR. Tests were sequential, not simultaneous. There are 64 measurement commands and 47 probe pairs; Ljubljana returned two probes instead of three.
What was calculated
Each probe’s ping mean is taken from the API’s statistics over the received replies. City summaries use the median of those means; the two-probe Ljubljana median is their midpoint. Packet totals sum the reported sent and received counts. MTR loss counts include only the final responding target.
What remains unmeasured
No IPv6, TCP, HTTP, throughput, sustained uptime or encrypted VPN tests are included. The archive does not record the Austrian VPS hardware specifications, workload or a valid CLI version string. Those details cannot be inferred from the Cyprus test machine.
What to do next
Use the permanent Looking Glass endpoints for future network measurements, recorded as a separate series. Repeat at different times, alternate target order and test from the actual networks using your service. A different endpoint can use a different route; do not merge its results with this VPS sample. Probe coordinates and city labels can be imperfect. Unusually low Munich RTTs and the high-variation Stockholm probe are retained, not treated as geographic proof.
Put the findings to work
Choose for your audience.
For Polish or Italian audiences
NDC2 is a sensible first candidate based on all sampled Warsaw and Milan probes. Verify the customer ISPs you actually serve.
For Vienna or Bratislava audiences
Interxion is a sensible first candidate for latency based on the sampled probes. Repeat the Bratislava tests to investigate the missing replies.
For diverse VPN customers
Offer or evaluate both locations where practical. Test encrypted throughput separately and avoid choosing from a single city median.
Infrastructure references: EDIS Global Vienna NDC2 and Vienna Interxion. Network results: supplied Globalping collection, 17 September 2026. Download the dataset.