2017 · Egypt · J - Information and communication · Nationwide network traffic hijacking and injection via ISP deep-packet-inspection equipment
Analysed 2026-11-07 by Hermes
Open this case in the tool →Nationwide network traffic hijacking and injection via ISP deep-packet-inspection equipment
No DV evidenced
Alternative: none defensible
(1) The expectation the framework would attach to a national communication network - safeguarding subscriber traffic integrity and assuring the third-party inspection platforms embedded in it - cannot be documented at any organisation here: the primary technical reporting attributes the injection to government-linked entities operating Sandvine PacketLogic DPI already installed on the state-owned network, so the network was the instrument of the operation rather than an organisation whose controls failed an external intrusion. (2) What the evidence shows: from at least October 2017, DPI middleboxes matching Citizen Lab's Sandvine PacketLogic fingerprint at a Telecom Egypt demarcation point hijacked Egyptian users' unencrypted web connections en masse and redirected them to affiliate ads and browser cryptocurrency mining scripts (the scheme Citizen Lab named AdHose, judged likely to be an effort to raise money covertly), alongside censorship via injected TCP resets; the campaign was multi-year in the historical OONI data, was discovered by outside researchers (ESET, OONI, Citizen Lab), and was never disclosed by the network or the vendor; in 2024 the U.S. Commerce Department blacklisted Sandvine because equipment it supplied to the government of Egypt enabled mass web monitoring, censorship and targeting of political actors and human rights activists. (3) Why this is Unknown rather than a DV-Type (atypical, attacker-side): the affected parties are Egyptian Internet users as a population, not an organisation with a security lifecycle, and no source places an unmet expectation inside any involved organisation - there is no documented origin (doctrinal or material) and break (leadership or execution) to run the operational test on; reporting confirms impact and attribution only. Under the method spec's decision rule for state operations without a victim organisation in the framework's sense, the DV-Type stays Unknown. Sandvine's own conduct is instructive but sits on the actor/vendor side: its February 2018 assertion that PacketLogic is 'not capable of Man-in-the-Middle (MITM) attacks and not capable of any form of payload injection' was contradicted by Citizen Lab's laboratory measurements of one-packet redirect injection, and the report concluded the company's 'strong safeguards' had 'come up short'. No lifecycle stage is used to derive a DV-Type; entry, detection and disclosure facts are recorded in Notes and the sources.
RM02 — Conflicting objective / incentive pressureRM44 — Public security-claim discrepancyAtypical, attacker-side state operation: DV-Type Unknown per the method spec's decision rule (no victim organisation in the framework's sense; the receiver is 'End user(s) / specially protected groups' - a diffuse population). The two coded risk markers describe the operator/vendor-side conditions the evidence shows (AdHose monetising subscribers' connections; Sandvine's capability denial contradicted by lab measurement and its safeguards found to come up short), recorded as instructive per the method spec's atypical-case rule, not as victim-organisation gaps. Attribution caveat on the row: the case file says 'The network company Sandvines injected spyware in the egyptian communication network and misused it to control the showcase of ads and to mine cryptocurency' and names Sandvine (state) as initiator; Citizen Lab instead attributed the injection behaviour to PacketLogic equipment 'likely by nation-states or ISPs' - operators with network access, most likely the ruling government per its press coverage - with Sandvine as the enabling vendor. The row's Egypt incident (affiliate ads + browser cryptocurrency mining, AdHose) matches the dataset's 01.10.2017 start date and Citizen Lab's injections observed since October 2017; targeted spyware injection with the same device class inside Egypt was documented later (2022-2023 Citizen Lab reports), while the 2018 spyware injection documented by Citizen Lab ran in Turkey/Syria. Sector J applies to the state telecom network infrastructure named in the row, though the harm fell on a diffuse end-user population; no sector of a single victim organisation strictly applies. The OONI 2016 Egypt interference post (first public detection of the redirects) was unreachable from this container and is therefore not cited; the Citizen Lab report itself builds on the OONI measurements. Detection stage (for the record): discovered by outside researchers only (ESET Sept 2017, Citizen Lab March 2018), no operator detection or disclosure - matching the dataset's third-party-disclosure field.