Exfiltration over Cloud Storage
Also called cloud exfil.
Stealing data by uploading it to a legitimate cloud service.
What it is
Moving stolen data out of a network by sending it to a widely used cloud storage or sharing service rather than to obviously attacker-owned infrastructure. The traffic blends in with sanctioned business use of the same platforms.
Why it matters
Because uploads to popular cloud providers are routine and often allowed, this exfiltration path hides in plain sight, so detection leans on data-volume baselines and which accounts or destinations are actually expected.
In practice
Noticing a server pushing gigabytes to a personal cloud account over HTTPS, the egress leg of a staged data theft.
What to look for
- Large uploads to personal cloud storage.
- Sync clients installed or run by unusual users.
- Uploads during off hours.
Common mistake
Allowing a service because the company uses it. Personal and company accounts share the same domains.
Practice it nearby
CyberKit has no tool for this term itself. These are the closest places to practice it.
- PCAP Triage
A first pass over a capture: conversations, protocol mix, DNS, HTTP hosts and TLS names.
- Network Case
A host suspected of beaconing: work its capture from scope to beacon, then the port recall and subnet math behind it.
In the field
Well-known tools where this shows up on the job. They are listed to show the landscape, not as a recommendation.
-
Amazon GuardDuty Cloud service
AWS threat detection that analyzes CloudTrail, VPC flow logs and DNS logs.
-
Microsoft Defender for Cloud Commercial
Posture and threat protection for Azure, AWS and Google Cloud.
-
Wiz Commercial
Maps cloud resources, identities and exposures into one graph of risk.
-
Prowler Open source
An open source tool that checks cloud accounts against benchmarks such as CIS.
Read next
- Data Staging: Collecting and bundling stolen data before it leaves the network.
- Exfiltration over DNS: Smuggling data out by encoding it into DNS queries.
- Encrypted Traffic Analysis: Inferring intent from encrypted traffic's shape and metadata.