2012 · J - Information and communication · State-sponsored espionage: multi-year data theft via complete network takeover
Analysed 2026-11-06 by Hermes
Open this case in the tool →State-sponsored espionage: multi-year data theft via complete network takeover
Expertise to Execution
Evidence coverage: 5/6 stages, weight 30/60
Decisive stage: containment (clearest evidence) · Alternative: Strategy to Execution
(1) Telecommunications providers are designated critical infrastructure holding call detail records and subscriber data; the expectation is that internet-facing services are hardened and minimally exposed, crown-jewel CDR/billing/domain-controller systems sit behind privilege tiering and segmentation, and exfiltration of subscriber data is visible to the operator's own monitoring. (2) In practice, in a campaign running since at least 2012 and first spotted by the external security vendor in 2018, a web shell on a vulnerable, publicly-facing IIS server gave the actor a foothold; it dumped credentials, moved laterally to production, database and billing servers, took full control of the Domain Controller, kept access through rogue high-privileged domain accounts, and queried and exfiltrated CDR, billing, e-mail and location data for years in small volumes that stayed under the radar, returning in four further waves over six months despite active remediation. (3) The unmet expectations were all known technical remedies with a material origin - patching and reduced exposure of public web servers, privilege tiering and directory hygiene, egress monitoring, complete eradication of persistence - and the chain broke in execution: they were never realised in the provider's architecture and operations. No source documents a strategic artefact that was not implemented (which would make it Strategy to Execution) or an expert warning leadership ignored (Expertise to Leadership), so the classification rests on the containment stage, where the primary forensic report most clearly shows material knowledge not executed; the entry vector is documented but not by itself load-bearing for the type. (4) RM30 (Expertise to Execution lens): containment could not stop the compromise because the architecture did not segregate crown-jewel systems, so one web-server intrusion ended in complete takeover of production, billing, database and domain-controller systems (causal role: Amplifying). RM41 (Expertise to Execution lens): after initial access the actor escalated to domain-wide control - credential dumping plus rogue high-privileged domain accounts reused between waves show directory hygiene and privilege tiering were not executed (causal role: Amplifying). RM27 (Expertise to Execution lens): the vulnerable, publicly-facing IIS server that carried the web shell stayed exposed without the recommended additional WAF layer and reduced internet exposure (causal role: Enabling). RM38 (Expertise to Execution lens): remediation waves stopped the attacker but did not eradicate its access, since the rogue accounts and persistence survived and the actor returned four more times within six months (causal role: Perpetuating).
RM27 — Configuration hardening / external exposureRM30 — Segmentation / containment weaknessRM41 — Privilege-escalation / directory-hygiene pathRM38 — Containment / eradication failure during responseEuRepoC row: no receiver_name; the victim is a global telecom sector (receiver_country 'Global (region)'), named descriptively; Country left null per the convention for diffuse multi-organisation campaigns. Year 2012 from eurepoc_extra.start_date (campaign active since at least 2012; the 2018 date in the row description is Cybereason's identification year). The forensic detail rests on one large telecom provider that was a Cybereason customer; the campaign hit several unnamed providers worldwide, so provider-specific lifecycle facts are the sample documented in the vendor report. The original Cybereason blog URL in the case file now redirects to a LevelBlue marketing page; cited instead is the 25-06-2019 Wayback snapshot opened during this analysis. The case file's other source URLs (Alloy Taurus, PingPull, Salt Typhoon) cover later, adjacent campaigns by related MSS-linked actors, not this intrusion, and were not used. Attribution to APT10/Tianjin MSS is by the vendor and the IT-security community (consistent with the 2018/2019 US charges against APT10 members), not a victim or regulator statement; the DV is assessed victim-side and the attribution does not affect it. The dataset description matches the case (quote of the Cybereason executive summary); no description/victim mismatch.