Yunshan offers DeepFlow, a zero‑code observability platform that uses eBPF and Wasm to automatically collect and correlate metrics, traces, logs, and profiling data across applications, containers, VMs, and network devices without modifying code. The platform provides full‑stack correlation and smart‑encoded storage for fast, low‑cost queries, while its NSP component adds SDN/NFV‑based network orchestration and AI‑driven diagnostics for cloud, edge, and 5G environments.
Funding
$11M raised to dateRaised to date based on public sources. This may differ from the amount the company actually raised and is based only on what is publicly available on the internet.
Founders
Product
Problem
Enterprises operating cloud‑native, micro‑service and large‑scale network environments lack a unified, low‑overhead way to collect, correlate and analyze performance data across applications, containers, virtual machines, and network devices. This leads to blind spots, manual instrumentation, and slow fault isolation, especially in AI model training, 5G core networks, and edge computing scenarios.
Solution
Yunshan delivers DeepFlow, a zero‑code, eBPF‑ and Wasm‑based observability platform that automatically captures metrics, traces, logs and profiling data from any process, container, VM or network element without modifying application code. Collected data are enriched with unified tags and stored using a SmartEncoding scheme that reduces storage costs while enabling fast, BigTable‑like queries. DeepFlow provides full‑stack correlation—from CPU/GPU function profiling to network packet analysis—allowing minute‑level fault localization and performance optimization. Complementary to DeepFlow, the NSP (Network Services Platform) offers SDN/NFV‑driven, template‑based network orchestration, automated service provisioning, and multi‑tenant management for data‑center, edge and WAN environments, integrating seamlessly with DeepFlow’s observability data to support automated network diagnostics and AI‑driven intelligent agents.
Target Audience
Primary customers are cloud service providers, telecom operators, large enterprises and SaaS vendors that run containerized micro‑services, AI training/inference workloads, or 5G/edge network infrastructures and require automated, code‑free observability and network service automation.
Features
- Zero‑intrusion data collection using eBPF across servers, containers, serverless pods, VMs and network devices
- Automatic distributed tracing and real‑time calculation of LLM metrics (TTFT, TPOT) without code changes
- Continuous function‑level profiling (CPU, GPU, memory) with flame‑graph visualizations and low (<1%) overhead
- SmartEncoding tag architecture that separates metadata from payload, achieving up to 10× lower storage than traditional ClickHouse schemas
- Integrated network performance monitoring (full‑stack path tracing, RDMA profiling, 100+ metrics) for 5G core, edge and hybrid cloud topologies
- SDN/NFV orchestration engine with TOSCA‑based service templates, multi‑vendor device abstraction, and self‑service portal
- AI‑powered observability agents that ingest collected data to provide automated diagnostics, predictive alerts and recommendation loops