Handling XML/SVG Sanitization for Large (10GB) Files in Concurrent TUS

Hey,

Hope you're keeping well.

For sanitizing a 10?GB XML/SVG in a concurrent TUS workflow, you need to avoid DOM-based parsing entirely and stream the content. Use XmlReader or System.Xml.XmlReader.Create with async streaming from S3 via the AWS SDK GetObjectAsync with ResponseStream) so you process and sanitize elements sequentially without loading the full file into memory. If uploads are chunked, sanitize each chunk independently before merging, and coordinate concurrency with a distributed lock (e.g., Redis RedLock) or a queue-based pipeline (Azure Storage Queue, Kafka) to ensure only one final merge pass. Disable DTD and external entity resolution XmlReaderSettings.DtdProcessing = Prohibit) to mitigate XML injection risks, and apply schema or whitelist-based filtering during stream processing. This approach keeps memory usage predictable while maintaining throughput for multiple concurrent users.

Thanks and regards,

Taz