Airlock Quick Start on Google Cloud

Airlock on Google Cloud is a virtual machine-based product. A free trial period is available during which there is no charge for the Airlock software but there may be charges for the underlying Google Cloud infrastructure.

Launch Airlock in Google Cloud

Go to Airlock in the Google Cloud Marketplace.
Click the Launch on Compute Engine button.

Virtual Machine Recommendations

The general purpose machine type is n2-standard-2 and this machine type should be adequate for most use-cases. We recommend 8 vCPUs and 8-16 GB of RAM for a production deployment.

Congratulations! You have deployed Airlock in Google Cloud. You are now ready to filter text!

Try it out!

With Airlock now running we can take it for a spin. We will send some text to Airlock and inspect at the response we get back. The Airlock virtual machine running in your cloud account should have a public IP address (unless you customized the deployment). We will use that public IP address to interact with Airlock.

Airlock, by default, will be configured with an HTTPS listener on port 8080 using a self-signed certificate. It is recommended that prior to use in a production environment the self-signed certificate is replaced by a valid certificate owned by your organization.

In the command below, replace <PUBLIC_IP> with the virtual machine’s public IP address or public host name.

curl -k -X POST https://<PUBLIC_IP>:8080/api/filter --data "George Washington was a patient and his SSN is 123-45-6789." -H "Content-type: text/plain"

With this command we are sending the text in the command to Airlock for filtering. Airlock will identify the patient name (George Washington) and the SSN (123-45-6789) and redact those values in the response. You can always use curl to send text to Airlock as in these examples but there are also SDKs you can use, too, to integrate Airlock with your applications.

Redacting Sensitive Information from Text

The types of sensitive information that Airlock identifies and removes is controlled by policies. By default, Airlock includes a filter profile that includes many of the types of sensitive information, such as names and social security numbers. We can send text to filter to Airlock for filtering using this default filter profile with the following command:

curl -k -X POST https://localhost:8080/api/filter -d @file.txt -H "Content-Type: text/plain"

This command sends the contents of the file file.txt to Airlock. Airlock will apply the enabled filters and return a plain-text response consisting of the filtered text. (Replace localhost with the IP address or host name of Airlock if you are not running the command where Airlock is running.) You can also send text directly in the request instead of sending it as a file:

curl -k -X POST https://localhost:8080/api/filter --data "Your text goes here..." -H "Content-type: text/plain"

Next Steps

Now that you have Airlock running and know how to send text to it you are ready to integrate Airlock into your existing workflow and systems. Airlock’s API details how to send files to Airlock. Clients for some languages for Airlock’s API are available on GitHub.

Example Uses

Here's a few examples showing how to use Airlock with some common big-data and streaming applications.

Description	Technologies
Remove sensitive information from text in an Apache NiFi dataflow	Apache NiFi
Remove sensitive information from text using AWS Lambda in an Amazon Kinesis Firehose pipeline	Amazon Kinesis, AWS Lambda
Removing PII/PHI from OpenAI ChatGPT API Requests	ChatGPT
Redacting PHI and PII in Apache Kafka Data streams	Apache Kafka

Last modified: 17 November 2023