URGENT: Singapore Network Specialist Wanted - Fixed Bounty for <300µs Latency to AS62014

₹600-5000 INR

Closed

Posted about 1 month ago

₹600-5000 INR

Paid on delivery

I am an infrastructure architect for a High-Frequency Trading (HFT) desk. I am looking for a server scout, network engineer, or local datacenter enthusiast to find a hosting provider in Singapore (VPS, Dedicated, Bare Metal, or Cloud) that offers significantly lower latency than my current setup. The Goal & The Bounty Target: I strictly need a server that delivers UNDER 300 microseconds (<0.3ms) stable latency to my target IPs. Payment: Fixed Bounty of ₹5,000 INR (~$60 USD). (Note: If you manage to find an extreme edge case like <50µs or <100µs, I am open to discussing a handsome bonus, but the primary deal is ₹5,000 for <300µs). I do not care about the server specs (1 Core / 4GB RAM is enough). It can be a cheap VPS, a premium dedicated server, or a local ISP. If you can prove the routing latency is under 300µs, you get paid. Target Network Details Target IPs (Telegram AS62014 - Singapore): [login to view URL] Metric: Round-Trip Time (RTT) in MICROSECONDS (µs). Technical Context & Our Research So Far (READ BEFORE BIDDING) Do not waste time testing random commodity VPS providers. We have been researching this for 1 months. Here is what you need to know to win this bounty: Current Benchmark: I already achieve 500 microseconds (~0.5ms) RTT using Kamatera (which utilizes Cogentco transit). Do not bid if you can only find 400µs or 500µs. I strictly need <300µs. The Physical Limit: Speed of light in fiber means round-trip latency is ~10.8 µs per km. Therefore, the maximum physical distance allowed to hit our target is strictly ~9-14 km. The server MUST be physically in the same building or campus as the target. The SG3 Wall: We tested premium providers in Equinix SG1/SG3 (like Global Secure Layer) who have active Private Network Interconnects (PNI) with Telegram. They bottomed out at 0.6ms due to physical fiber distance to the target and Anycast routing. The Scrubbing Layer: MTRs usually show 4 hidden hops (* * * *) before reaching the target. This is an Anycast scrubbing firewall. This processing adds ~0.2ms. Your chosen network path must bypass this or be physically close enough to absorb this delay. Our Hypothesis: We strongly suspect the target is physically located in or very near Equinix SG5 (Tanjong Kling), or strictly requires a Tier-1 upstream like Cogentco that has a direct, massive pipe there without hairpin routing. What You Must Deliver To claim the payment, you must provide: The Provider Name: Which hosting company or ISP provides this latency? Proof of Latency (Choose one): Option A (Best): A test IP address so I can verify the ping/MTR from my end. Option B: A video recording of the Linux terminal showing ping and mtr to the target with a visible timestamp. Option C: Temporary SSH access to the instance for me to verify. Cost: The monthly rental price of the server (I will pay for the server rent myself; your bounty is separate). Verification & Rules Strictly NO Advance: Do not ask for upfront payment to "search". This is a payment for a verified result. No Fake Results: Do not use fake ICMP responders or altered terminal outputs. I am a systems engineer and will verify the BGP path using my own tools. Stability: The latency must be consistently under 300µs (low jitter), not just one lucky packet. Apply ONLY if you have access to various Singapore data centers to test, or if you know the local ISP/BGP infrastructure intimately.

Network Engineering

Cloud Computing

Network Monitoring

Telecommunications Engineering

Project ID: 40317016

About the project

3 proposals

Remote project

Active 27 days ago

Looking to make some money?

Email address

Benefits of bidding on Freelancer

Set your budget and timeframe

Get paid for your work

Outline your proposal

It's free to sign up and bid on jobs

3 freelancers are bidding on average ₹4,100 INR for this job

@shaxzodj

Hi, This is a very interesting and challenging task — I understand you're operating close to the physical limits of latency. Instead of making unrealistic promises, I’d like to approach this strategically. My plan: - Test multiple Singapore-based providers (including lesser-known/local ISPs) - Use looking glass servers and remote instances to analyze routing paths - Focus specifically on networks near Equinix SG5 / low-hop Tier-1 routes - Identify paths that minimize scrubbing/anycast delay I will only claim the bounty if I can consistently achieve <300µs RTT as required. If reaching <300µs is not possible, I will still provide you with the best achievable latency, full MTR analysis, and insights into routing limitations — which may still be valuable for your HFT setup. No upfront payment needed — fully result-based. Let me know if you'd like me to proceed. Thanks

₹5,000 INR in 4 days

0.0

(0 reviews)

0.0

@TheGrayistVA

Hi, I am a Systems Engineer specializing in hardware optimization and network infrastructure. I’ve reviewed your 500µs benchmark and the SG1/SG3 "Wall." At <300µs, we are solving for a "Zero-Hop" or "Same-Campus" interconnect. Initial Research & High-Probability Leads: Based on current BGP/PeeringDB data, Telegram (AS62014) has a 600Gbps presence at Equinix IX and SGIX. Your 500µs via Kamatera (Cogent) likely suffers from a metro-fiber hairpin. I have identified three primary targets to bypass this: Leaseweb (SG5/SG2): Known for flat routing and proximity to the Equinix West campus, minimizing metro-fiber latency. G-Core S.A. (AS199524): A direct peer of Telegram in SG. Their edge-focused routing often bypasses the 4-hop Anycast scrubbing layer. Cherry Servers (Equinix SG5): Bare Metal in the Tanjong Kling facility. If the target is indeed SG5, this provides the shortest possible physical fiber run. My Approach: Zero-Hop Verification: Prioritizing providers with Direct PNI or shared L2 fabric to stay under the ~10km physical fiber limit. Jitter Analysis: I will provide 100-packet MTR traces to ensure stability (Low StdDev) for HFT requirements. L2/Bare-Metal: Avoiding SDN-heavy cloud providers to eliminate hypervisor jitter. Delivery: Provider name and specific DC Campus. Verified MTR/Ping proof with timestamps. Monthly rental cost. I operate on a result-first basis; no upfront payment required. I am ready to begin testing immediately. Best regards,

₹2,800 INR in 7 days

0.0

(0 reviews)

0.0

@amrishv

Hi, This is an interesting and realistic challenge — and I appreciate the clarity in your requirements. You are absolutely right about the physical and routing constraints. Achieving <300µs RTT essentially requires: - Same facility or immediate proximity (≤10 km fiber path) - Minimal hops / direct peering (preferably PNI) - Avoiding Anycast scrubbing layers Instead of testing generic VPS providers, I would approach this in a targeted way: 1. Identify infrastructure inside/near Telegram’s Singapore presence (likely Equinix SG or similar IX-heavy facilities) 2. Evaluate providers offering: - Direct IX participation (SGIX, Equinix Fabric, etc.) - Low-hop routing (no unnecessary transit layers) 3. Run controlled latency benchmarks (hping/tcp ping, not just ICMP) 4. Compare routing paths and eliminate scrubbing/Anycast-heavy routes I work extensively with VPS and cloud infrastructure and can provision/test across multiple providers quickly, including custom routing checks and latency validation. Additionally, if required, I can also provide a dedicated VPS setup optimized specifically for ultra-low latency workloads. Before proceeding, I would confirm: - Whether UDP/TCP latency is acceptable vs ICMP - If testing from multiple ASN paths is allowed I can start immediately and focus only on solutions that realistically meet your <300µs requirement. Thanks, Amrish

₹4,500 INR in 7 days