What was exposed
Replaying the TrueDialog November 2019 exposure
- 1
Configured: the search database answers the internet
not statedThe Elasticsearch database on Azure is left reachable from the internet with no protection; the company names a non-published network port, and the setting, who set it and when are not stated.
- 2
Reachable: read from a browser with no credential
not stated to Nov 29, 2019Anyone on the internet can read the messages and the stored logins with no credential; the finders read the schema and the entries from a browser.
- 3
Found and closed: found Nov 26, closed Nov 29
Nov 26 to 29, 2019vpnMentor researchers find the database by port scanning and contact TrueDialog, TechCrunch contacts the company too, and the database is closed on November 29.
604 GB and nearly 1 billion entries per the finders: tens of millions of SMS messages with recipients' details, millions of customer account logins in cleartext or base64, user details and technical logs. Open from a date not stated until November 29, 2019, three days after the finders found it; the company calls the window short and says it found no evidence that anyone other than the reporting analyst downloaded or viewed the data.
Where cloud-provider controls fit
An Azure Policy Deny on inbound network security group rules from every internet source bounds the opening on any port where such a rule was the opening, and the sources do not name the setting that opened the database. With the opening in place, an Azure Virtual Network Manager security admin rule that denies internet traffic would still have kept readers on the internet out. On AWS, an OpenSearch Service domain with VPC access has no public endpoint to open where the database is such a domain, and on Google Cloud a custom constraint bounds the firewall rule that would open it; a service control policy on internet gateways and an OCI Security Zone remove the internet path instead. Whether the database asks for a password and how your application stores its customers' passwords stay with the database software and the application, and the scan that found it and the fix that closed it stay outside the provider's controls.
Could this have been prevented?
In hindsight, no control would have stopped it. 2 controls would have limited how far it went. 3 steps have no cloud-provider control. The counts below are control assessments, not attack stages: several can address the same step. The same replay with the controls in place before the incident shows where each phase would have been refused or bounded.
- 0
- Prevented
- 2
- Contained
- 3
- Not covered
How it could have been prevented
The same replay with the controls in place. Green marks where a step is refused, blue where it is bounded, and red where no control reaches it. Open a row to read the control, why it acts, and its caveat.
Replaying the TrueDialog November 2019 exposure
1 · Configured: the search database answers the internet
1 Contained4 on other clouds2 · Reachable: read from a browser with no credential
1 Contained1 Not covered3 on other clouds3 · Found and closed: found Nov 26, closed Nov 29
2 Not coveredKey facts
- Per the finders, the database held 604 GB and nearly 1 billion entries at their last look
- The records held tens of millions of SMS message entries with names, phone numbers, email addresses and message content; TechCrunch, which examined a portion of the data, saw two-factor and password reset codes among the messages
- The records held millions of account logins of TrueDialog's customers, some passwords in cleartext and others base64-encoded, per the finders
- vpnMentor researchers found the database on November 26, 2019 by port scanning and contacted TrueDialog on November 28; when the exposure started was not stated
- The database was closed on November 29, 2019, per the finders; TechCrunch said its own contact led TrueDialog to pull it offline
- TrueDialog said the data sat at a non-published network port, now secured, for a short period, and that it found no evidence anyone other than the reporting analyst downloaded or viewed it
Sources
Every fact in this breakdown comes from the public reporting below. The replay is a model of that reporting, not a reconstruction of the environment it describes.
- 1.Report: Millions of Americans at Risk After Huge Data and SMS LeakvpnMentor · Jul 17, 2023 · ResearchAzure hosting, Elasticsearch, no protection, browser access, size, contents, cleartext and encoded credentials, discovery method, notification date, closure date, owner attribution, the finders' advice
- 2.Millions of SMS messages exposed in database security lapseTechCrunch · Dec 1, 2019 · Newsindependent check of the contents, no password, years of messages, takedown after contact, company silence
- 3.Welcome back from the holiday, Americans! Here's who leaked data while you were awayThe Register · Dec 2, 2019 · Newscompany statement, notification day, exposure window as the company states it, network port, no evidence of other access, Azure hosting
The estate in the replay is illustrative and drawn at the level of the exposure; the owner's real environment is not public. Verdicts describe what a control does when it is in place before the data is exposed. They are not an assessment of the owner's security program.