Free test available for France , UK or SG on Telegram Join Telegram
Performance Benchmarks

Proxy Type Speed Comparison

Comprehensive performance benchmarking: evaluating round-trip latency, throughput ceilings, TLS handshake duration, and job completion speed across all proxy architectures.

PXM2 Lab September 2026 8 min read
<5ms Datacenter ping
15–40ms 5G mobile latency
150Mbps+ 5G carrier throughput
7+ Testing nodes
  • Round-trip latency across physical network mediums — benchmarking cloud fiber backbones, DSL/cable loops, and cellular radio links.
  • Throughput and bandwidth saturation limits — evaluating 10Gbps datacenter uplinks versus 4G LTE and 5G NR mobile speeds.
  • Connection setup and TLS handshake overhead — analyzing TCP SYN/ACK, TLS 1.3 key exchange, and HTTP CONNECT negotiation.
  • Real-world job completion velocity — why lower ban rates often make mobile proxies faster end-to-end despite radio latency.
Network Benchmarks Throughput Audit
Datacenter latencySub-5ms fiber cross-connect
Residential latency40–120ms (variable P2P hops)
Static ISP latency10–25ms (fiber/coaxial link)
Mobile 4G/5G latency15–45ms (cellular base station)
Carrier 5G NR Acceleration

Direct cellular tower interconnects delivering fiber-grade throughput.

Effective Scraping Velocity

Eliminate slow retry loops and rate-limit backoff penalties.

Proxy speed is frequently measured using simplistic ICMP ping tests that fail to reflect real-world application performance. While a datacenter proxy hosted in a cloud facility can deliver a 5ms ICMP ping, its effective crawling throughput plummets to zero the moment anti-bot security systems challenge it with JavaScript puzzles or CAPTCHAs.

True network speed encompasses four distinct engineering metrics: Round-Trip Time (RTT), TLS negotiation latency, Time to First Byte (TTFB), and sustained TCP bandwidth throughput. Understanding how cellular radio modulation (4G LTE and 5G NR) impacts these stages enables engineers to optimize automation pipelines for maximum velocity.

This benchmark study breaks down network physics across datacenter, static ISP, residential P2P, and mobile proxy tiers with real-world telemetry.

Deconstructing Proxy Latency: RTT, Radio Links, and TLS Overhead

When an HTTP client executes a request through an intermediary proxy, total latency is the sum of four discrete network phases:

  1. Hop 1: Client to Proxy Gateway (5ms–20ms)

    The transit time from your automation server to the proxy entry node over high-speed fiber backbones. Minimizing physical distance to the proxy gateway optimizes this stage.

  2. Hop 2: Radio Interface Handshake (15ms–35ms)

    For mobile proxies, packets traverse the cellular air interface (LTE eNodeB or 5G gNodeB base station). Radio scheduling frames (TTI) introduce an irreducible 10ms–25ms radio modulation latency.

  3. Hop 3: TLS 1.3 Cryptographic Handshake (1 RTT)

    Negotiating cryptographic cipher suites and exchanging ephemeral Diffie-Hellman keys requires one full round trip between the proxy exit node and the target server.

  4. Hop 4: Target Server TTFB & Content Streaming

    Server-side application processing time plus TCP sliding window bandwidth throughput for streaming HTML, JSON, or media payloads back to the client.

Cellular Radio Physics: 4G LTE Cat 6 vs 5G NR Modulation

Cellular proxy performance is governed by wireless radio modulation standards. The table below details the physical differences between 4G LTE and 5G NR modem hardware:

4G LTE Advanced (Cat 6) Sub-6 GHz

Carrier aggregation combining two 20 MHz LTE bands with 64-QAM / 256-QAM downlink modulation.

  • Radio Interface RTT: 25ms to 45ms
  • Downlink Throughput: 40 Mbps to 120 Mbps
  • Uplink Throughput: 15 Mbps to 35 Mbps
5G NR (Sub-6 / C-Band) Ultra-Low Latency

Next-generation New Radio standalone (SA) and non-standalone (NSA) modulation with flexible subcarrier spacing.

  • Radio Interface RTT: 12ms to 22ms
  • Downlink Throughput: 150 Mbps to 450 Mbps
  • Uplink Throughput: 40 Mbps to 90 Mbps

Synthetic RTT vs Effective Crawling Throughput Reality

A low ping is useless if your requests trigger CAPTCHA challenges. Consider this real-world benchmark when scraping 1,000 pages from a Cloudflare-protected e-commerce portal:

The Effective Velocity Paradox: Datacenter proxies boast a 10ms synthetic ping, but with an 85 per cent Cloudflare block rate and 15-second CAPTCHA solve delays, completing 1,000 pages takes 42 minutes. Dedicated mobile proxies exhibit a 35ms radio ping, but with a 0.1 per cent block rate, those same 1,000 pages finish in just 4.5 minutes. Real speed is measured in successful pages per minute.

Benchmark Matrix: Speed & Latency Across Four Architectures

Performance Vector Datacenter Static ISP P2P Residential Dedicated 5G Mobile
Raw RTT Ping Latency 5ms – 15ms 15ms – 30ms 75ms – 180ms 18ms – 35ms
TLS 1.3 Handshake Duration 15ms – 25ms 25ms – 40ms 90ms – 220ms 30ms – 50ms
Time to First Byte (TTFB) 45ms 65ms 240ms 85ms
Bandwidth Throughput 500–1,000 Mbps 100–300 Mbps 5–25 Mbps (Peer bottleneck) 150–450 Mbps
Jitter / Variance Under 2ms Under 5ms 50ms – 120ms (Erratic) 5ms – 12ms
Effective Crawl Velocity (WAF) Very Slow (Blocked) Moderate Slow (Peer drops) Fastest (Zero Blocks)

4-Hop Network Latency Waterfall: TTFB Breakdown by Architecture

Raw ICMP ping times provide a deceptive picture of proxy performance. In production HTTP/HTTPS environments, total Time to First Byte (TTFB) is dominated by the physical transit across four network hops: client-to-gateway transit, TLS negotiation, cellular radio scheduling, and upstream target round-trip time. The waterfall benchmark below illustrates where latency actually accumulates across proxy tiers:

Network Telemetry

Time to First Byte (TTFB) Component Breakdown (Milliseconds)

Benchmarking HTTPS GET request to Cloudflare-edge origin across proxy connection tiers

Hop 1: Client to Gateway TCP
Hop 2: TLS 1.3 Handshake
Hop 3: Last-Mile Radio / Local Loop
Hop 4: Target Origin Processing
0 ms 130 ms 260 ms 390 ms 520 ms 650 ms Datacenter Fiber 5G Mobile (Sub-6) 4G LTE Cat 6 Modem P2P Residential Peer 48 ms 92 ms 146 ms 480 ms (Peer Jitter)
Datacenter TTFB 48 ms Lowest latency (92% ban rate)
5G Mobile TTFB 92 ms Sub-100ms with 99.4% trust
4G LTE TTFB 146 ms Predictable radio latency
Residential Peer TTFB 480 ms High bufferbloat & WiFi jitter

While datacenter proxies register lower raw synthetic ping, their instant ban rates on bot management systems mean effective extraction throughput collapses. 5G NR and 4G LTE mobile proxies provide the optimal balance: sub-150ms TTFB with total immunity against IP reputation blocking.

TCP Congestion Control: BBR vs Cubic on Cellular Links

Wireless cellular links exhibit dynamic packet delay variation and transient packet loss that cause traditional TCP Cubic congestion algorithms to prematurely throttle transmission speeds. Google's BBR (Bottleneck Bandwidth and RTT) algorithm, in contrast, models real-time radio pipe capacity rather than treating packet loss as a congestion indicator.

PXM2 proxy gateways optimize transport pipelines by employing TCP BBR congestion control across cellular radio modems. This maximizes throughput over 4G LTE and 5G connections, ensuring that high-volume scraping and data streaming pipelines achieve peak bandwidth utilization even in fluctuating cellular environments.

Test Fast Mobile Proxies

Access low-latency 4G and 5G cellular modems across our global node infrastructure:

🇫🇷

France

3 Operators 20-100 Mbps
Starting from
$4.34 for 1 hour
4G
Available Operators:
Orange Bouygues SFR
🇮🇳

India

3 Operators 20-30 Mbps
Starting from
$2.74 for 1 hour
4G
Available Operators:
Airtel Jio Vodafone Idea (Vi)
🇵🇱

Poland

1 Operator 20-80 Mbps
Starting from
$3.99 for 1 hour
4G
Available Operators:
Play
View all locations →

Frequently Asked Questions

Which proxy type offers the fastest raw ping and lowest latency?

Datacenter proxies provide the absolute fastest latency, typically sub-5ms to major cloud providers and tier-1 exchanges. They operate on enterprise optical fiber backbones directly connected to major internet exchange points (IXPs).

How does cellular radio latency affect 4G and 5G mobile proxy speeds?

4G LTE cellular connections introduce a physical radio over-the-air hop that typically adds 25ms to 50ms of base latency. 5G NR (New Radio) architectures dramatically reduce this air interface overhead, achieving latencies between 12ms and 25ms under optimal signal conditions.

Why do peer-to-peer residential proxies suffer from high latency variance?

P2P residential proxies route traffic through consumer smartphones and home Wi-Fi routers. Variable Wi-Fi interference, consumer upload congestion, geographic distance from relay nodes, and underpowered host hardware cause extreme latency jitter and unpredictable spikes.

How does HTTP/2 and HTTP/3 multiplexing perform across different proxy tiers?

Datacenter and static ISP proxies excel at multiplexing dozens of concurrent streams over a single TCP connection due to zero packet loss. On mobile proxies, TCP packet loss on weak cellular signals can trigger head-of-line blocking; using HTTP/3 (QUIC over UDP) via modern mobile proxies mitigates this radio phenomenon.

What is effective throughput versus nominal connection speed in web scraping?

Nominal speed measures raw pipe capacity (e.g., 100 Mbps). Effective throughput measures successful, unblocked responses per minute. A 1 Gbps datacenter proxy suffering an 80 per cent block rate achieves lower effective throughput than a 30 Mbps mobile proxy with a 99 per cent success rate.

Experience Ultra-Fast 5G Mobile Proxies

High-speed 5G NR cellular modems, low radio latency, and unmetered gigabit backbones. Test a fast proxy today.

Deploy Fast Proxies