Customer Case Study · 80-Person R&D Company: Data Center Hosting + Encrypted Three-Site Interconnect + Full Desktop IT Support

Client Profile

This is a technology company focused on enterprise SaaS development, headquartered in Wangjing, Beijing, with R&D centers in Zhangjiang, Shanghai and Nanshan, Shenzhen, and roughly 80 employees in total. Teams across the three sites share a single code repository, generating more than 200 Git commits a day on average — you can imagine how intense the collaboration is. When CTO Mr. Li talked with us, he said product iteration speed had always been the company’s core competitive advantage, but over the past two years IT problems had held it back considerably — not because development itself was slow, but because the “infrastructure” kept failing at critical moments.

Before the Engagement: 3 Critical IT Problems

In Q3 2019 there was an unforgettable lesson. At 4 p.m. on a Friday, the RAID array on a Dell PowerEdge R740 server in the server room raised an alarm — a drive was on the verge of failure. The colleague handling IT operations on a “part-time” basis was already driving on the highway and said on the phone that he could not be back until Monday. More than 60 hours of business interruption followed: R&D ground to a halt, customers complained, and service was not restored until Monday afternoon. This was not an isolated incident — only the tip of the iceberg. During the remote-work period it got even worse. The Shanghai team reported that latency when connecting back to the Beijing code repository jumped from a normal 30ms to 200ms+, and half of all Git pushes failing and aborting became the norm. Over one month, VPN disconnections averaged 5 times per week, with each emergency fix taking about 40 minutes on average. R&D productivity was being eroded, and Mr. Li was too anxious to sleep. IT manager Lao Zhang did the math for me: at least 10 hours every week went into handling sudden IT issues — server alerts, blue-screened employee laptops, dropped network connections, printer failures — leaving no time at all for system optimization or security hardening. They had brought in two on-site outsourcing staff; response was slow, and when something went wrong they blamed each other, so in the end Lao Zhang had to cover for them himself.

What We Did

Our first meeting was in September 2020, and Mr. Li got straight to the point: “What I want isn’t an outsourcing vendor — it’s an IT partner who will fight alongside me.” I still remember that line today.

1. Week 1: Take stock of everything and deploy monitoring — We spent a week compiling an inventory of all network equipment across the three sites, including 2 Cisco Catalyst 9300 core switches, 1 Huawei USG6330 firewall and 6 servers. We deployed a Zabbix + Grafana monitoring platform, putting the status of every device on one screen. We also signed a Network Equipment Package agreement, formally activating a 7×14 non-working-hours response SLA.
2. Weeks 2–3: Replace the traditional VPN with sTrust Zero Trust — This was the turning point of the entire plan. We deployed an sTrust Zero Trust gateway in the Beijing server room, replacing the existing OpenVPN cluster. The three R&D centers were interconnected through encrypted tunnels, and measured cross-site latency dropped from 200ms to under 30ms. Developers said code sync felt “as smooth as a local disk.” The sTrust Zero Trust solution not only solved the connection stability problem but also enabled fine-grained access control — the Shanghai team could reach only the code repository and CI/CD platform, and had no access to the finance system.
3. Week 4: Full desktop operations coverage — We posted an “IT support QR code” next to every employee’s workstation; scanning it joins a dedicated service group with a guaranteed response within 5 minutes. In the first month we handled 127 endpoint tickets, with an average response time of 4 minutes 32 seconds. Lao Zhang said employees now raise IT issues less often — not because there are fewer problems, but because everyone knows someone is on it, so small issues no longer fester into big ones.
4. Month 3: Off-site backup for the code repository — We deployed a backup node in the Shanghai server room in an active-active Beijing–Shanghai architecture, RPO=0, RTO<15 minutes. During an electrical maintenance window at the Beijing facility last November, the code repository failed over to the Shanghai node in 8 seconds, with no impact on R&D whatsoever.

The Results After One Year

This is my favorite part to present — the numbers don’t lie.

Metric	Before	After	Improvement
Cross-site collaboration latency	200ms+	30ms	↓ 85%
Branch VPN disconnections	5 times/week	0	↓ 100%
Time the IT manager spent leading incident response	10h/week	1h/week	↓ 90%
R&D productivity	Baseline	+12%	↑ 12%

The change in R&D director Lao Zhang is the most convincing. His first words at every morning meeting used to be “the servers went off again last night”; now his slides are full of product roadmaps and customer feedback. “I finally have time to think about technical architecture,” he says — and when he says it, there’s a light in his eyes.

What the Client Says

Mr. Li later referred two peer companies to us. One was a smart hardware company in Shunyi, the other a game studio in Yizhuang. In his words: “Choosing an IT provider is like choosing a business partner — reliability matters more than price.”

Who This Fits Best

This solution is no cure-all, but if you match the characteristics below, you will most likely benefit:

1. R&D-driven technology companies, 50–200 employees, whose code assets are their core lifeline
2. Multiple office locations, a headquarters plus 1–2 branch sites, with cross-city collaboration as the norm
3. No dedicated IT staff, or just 1–2 IT people, juggling multiple roles and constantly stretched thin
4. Long frustrated by VPN, with remote access lag, dropped connections, and high latency
5. Concerned about data security, with code and business data stored in scattered locations and no unified protection
6. Opaque IT budgets, always feeling money is being wasted without knowing where it goes

Start the Conversation