netcup Official Speedtest, Looking Glass & Datacenter Selection Guide

Table of Contents9 sections

Before deploying production workloads on any VPS or Root Server, running a quick ping and mtr route trace from your own network or target user region is the most reliable way to pick the right datacenter.

netcup operates infrastructure across five global locations spanning Central & Western Europe, the US East Coast, and Southeast Asia. This guide brings together the official Looking Glass tools, verified IPv4/IPv6 test IPs, CLI network diagnostic commands, and real-world latency benchmarks across Europe/UK, North & South America, and Asia-Pacific/Oceania.

---

1. Regional Routing Architecture: How the 5 Datacenters Compare#

Understanding how netcup (and its parent infrastructure group Anexia, AS197540 / AS42473) peers at major Internet Exchanges helps you choose the optimal location:

1. Europe & United Kingdom (Nuremberg, Vienna, Amsterdam — 5 to 20 ms):
   • Nuremberg (Germany) is netcup's flagship primary datacenter. It connects directly via redundant dark-fiber rings to DE-CIX Frankfurt (one of the world's largest Internet exchanges), delivering 3–10 ms within Germany/Central Europe and 12–18 ms to London/UK.
   • Vienna (Austria) connects directly to the Vienna Internet Exchange (VIX), making it the premier choice for Austria, Switzerland, Italy, and Central/Eastern Europe (2–15 ms).
   • Amsterdam (Netherlands) sits at the AMS-IX hub, offering ultra-low latency (4–12 ms) to the Benelux region, the UK, and Nordic countries, as well as direct transatlantic subsea cables.
2. North & South America (Manassas, Virginia / US-East — 10 to 65 ms):
   • Located in the Northern Virginia / Ashburn data center corridor (the primary internet hub of North America), netcup's Manassas facility delivers 5–20 ms to New York, Washington D.C., Atlanta, Toronto, and Chicago, 55–65 ms to US West Coast cities (Los Angeles, Seattle, Silicon Valley), and 70–95 ms across the Atlantic to London and Frankfurt.
3. Asia-Pacific & Oceania (Singapore & Eurasia Routes):
   • Singapore connects via international Tier-1 BGP transit in Southeast Asia, providing 2–15 ms within Singapore/Malaysia, 25–45 ms to Indonesia/Thailand/Vietnam/Hong Kong, and 55–85 ms to Tokyo, Sydney, and Perth.
   • Important Routing Note for East Asia: Because netcup's Singapore facility uses standard international BGP transit rather than costly China-premium private circuits (like CN2 GIA), certain mainland Asian consumer ISPs may route Singapore traffic across the Pacific via US-West. Always verify with mtr first—for many Eurasian connections, Nuremberg (160–210 ms direct via Frankfurt) provides more predictable routing and has no location surcharge.

---

2. Official Speedtest Server & Looking Glass#

• Official Nuremberg Speedtest Node: speedtest.netcup.de (official web speedtest & iperf3 endpoint for testing raw download/upload throughput);
• Official Looking Glass (LG): Accessible via the netcup help center and network status portal, allowing you to execute bidirectional ping, traceroute, and BGP route queries directly from netcup's core routers to your IP.

---

3. Public IPv4 & IPv6 Test IPs for All Datacenters#

Run ping and mtr from your local terminal against these verified test IPs to inspect real-time latency and hop-by-hop routing:

---

4. The 3-Step Network Benchmark Method#

Step 1: Measure Latency & Packet Loss with ping#

# Send 50 ICMP echo requests from your terminal
ping -c 50 202.61.254.56

• What to look for: Ignore single-packet min spikes and focus on avg (average RTT), mdev (jitter), and packet loss during peak evening hours.
• Healthy baselines:
  • Within Europe / UK to Nuremberg or Vienna: 5 ms – 20 ms, 0% packet loss.
  • Within North America to Manassas (US-East): 10 ms – 30 ms (East/Central) or 55 ms – 68 ms (West Coast), 0% packet loss.
  • Within APAC / Australia to Singapore: 15 ms – 65 ms, 0% packet loss.

Step 2: Trace Every Hop & ASN with mtr#

# Run 50 cycles in report mode showing Autonomous System Numbers (ASN)
mtr -rwzbc 50 202.61.254.56

• How to read the output:
  • Focus on the final destination hop: Only packet loss that persists all the way to the final hop indicates real network congestion.
  • Ignore intermediate hop ICMP rate-limiting: If hop 5 or hop 8 shows 50%–100% loss while the final hop shows 0.0% loss, that core router simply deprioritizes ICMP responses—actual TCP/UDP traffic is unaffected.

Step 3: Test Multi-Stream Throughput with iperf3#

# Test download throughput from netcup's speedtest server using 8 parallel streams
iperf3 -c speedtest.netcup.net -P 8 -R

• Why -P 8 matters: On long-haul links with 60–200 ms RTT (such as transatlantic US-to-Germany or Eurasia routes), a single TCP stream is limited by bandwidth-delay product (BDP) window scaling. Running -P 8 (8 parallel streams) alongside Linux TCP BBR reveals the true 2.5 Gbps virtual port capacity.

---

5. Global Regional Latency & Routing Matrix#

---

6. Final Datacenter Recommendations by Use Case#

1. For European & UK Users, SaaS Apps, or Global Best-Value Hosting:

Choose Nuremberg (Germany). It has zero location surcharge, the largest hardware inventory, the fastest provisioning, and first access to promotional offers. Choose Vienna for Austria/CEE or Amsterdam for UK/Benelux edge workloads.

1. For US, Canadian & Latin American Audiences:

Choose Manassas, Virginia (US-East). Sitting right next to Ashburn, it gives you dedicated AMD EPYC hardware in the heart of the US East Coast at European pricing.

1. For Southeast Asia, Australia & APAC Regional Workloads:

Choose Singapore, and run a quick mtr check from your target ISP beforehand to confirm direct regional peering.