2021 · Philippines · J - Information and communication · Distributed denial-of-service (DDoS) campaign against public websites
Analysed 2026-11-07 by Hermes
Open this case in the tool →Distributed denial-of-service (DDoS) campaign against public websites
No DV evidenced
Alternative: none defensible
(1) The victims are red-tagged independent publishers that had been the target of a massive DDoS campaign before (December 2018 - February 2019, later traced to two local ICT companies); their strategic answer was to run all three websites through Qurium Media Foundation, a host specialising in the defence of human-rights media, with forensic-grade monitoring, DDoS mitigation and attribution capability that none of the outlets could fund in-house. (2) In practice, floods on 17, 18 and 20 May, 6 and 16 June and 22-23 June 2021 were each logged within minutes, fingerprinted (HTTP floods, ApacheBench, concurrent Sn1per/Nikto scans) and attributed by the host; outages were brief - the worst being several hours on 22-23 June, with no lasting damage and no data theft - and Bulatlat's own account states Qurium 'effectively mitigated' the 'brief but frequent' attacks. Disclosure was prompt and victim-side: media statement on 22 June, forensic report on 22 June/1 July, public demands for a DOST/DICT investigation. (3) This is a confirmed negative/control case: the doctrinal-material alignment is positively evidenced, not merely unchallenged - the very control the outlets had built after 2019 (specialist-host mitigation with forensics) is what held, and in the sustained follow-on wave Karapatan ran with zero downtime against a 30,000-bot flood. The residual risk that materialised (a state-linked actor with botnet capacity) produced short disruptions, weighted intensity 1 in EuRepoC, rather than the disproportionate impact that the interaction of a knowing-doing gap and a doctrinal-material misalignment would produce; the entry mechanism is a volumetric flood of public websites and is not load-bearing for any victim-side gap. Two conditions the evidence does show are carried as markers without a DV lens: the availability protection could not fully hold during the largest flood (RM32), and the outlets had no in-house capability to mitigate or attribute such attacks (RM08) - both were compensated by the outsourced arrangement, which is why neither compounds into a dual vulnerability. No regulator or court examined the victim's controls; the reading rests on the host's forensic record and the victims' own disclosures. (Alternative readings were rejected: the repeat victimisation since 2018 was not coded as a learning failure because the documented change of mitigation posture demonstrably reduced impact.)
RM32 — Capacity / availability-protection gapRM08 — Security resourcing / capability constraintAttribution is one-sided: it rests on Qurium's forensic reports, and Qurium is the victims' hosting provider, i.e. victim-aligned. DOST denied involvement, said it leases parts of its IP space to other agencies, and by 6 July had named neither the tenant agency nor produced the promised investigation (Qurium had no other abuse channel - its report email to DOST bounced). The zero-downtime evidence refers to the sustained Karapatan wave of July-August 2021 (Qurium press release of 18 August 2021), which follows the row's end date of 23.06.2021 and is treated only as corroboration of the mitigation arrangement, not as part of the incident. The earlier comparable campaign (December 2018 - February 2019, legal case against two facilitating ICT companies) is documented in the Qurium press release of 1 July 2021; not coded as RM21 because the evidence shows lessons were turned into a validated change (host-based forensic mitigation), with impact dropping from weeks of downtime to hours. Data quality: the row carries two receiver entries (Social groups; Media) with receiver_country Philippines for both; no receiver_name is given, so the victim is named from the description's three organisations. Case file carries no related_rows. Unattended run: no action requiring consent was needed. All cited URLs were opened and read during this analysis (07-11-26).