end-to-end-encryption.md

Protect result

In previous tutorials, we saw how to build Confidential Computing applications that run securely inside enclaves and combine them with confidential assets to get the most out of confidential computing advantages. In this chapter, we will push things further to protect the workflow in an end to end mode. That means the next step would be encrypting results.

{% hint style="warning" %}

Before going any further, make sure you managed to Build your first application with Scone framework.

{% endhint %}

{% hint style="success" %}

Prerequisites:

• Docker 17.05 or higher on the daemon and client.
• iExec SDK 8.0.0 or higher. Install the iExec SDK
• Familiarity with the basic concepts of Intel® SGX and SCONE framework.

{% endhint %}

{% hint style="info" %}

You don't need to change your application's code or redeploy it to add this feature.

{% endhint %}

Assuming your application is deployed (if not please check how to do it with Scone), before triggering an execution you need to generate an RSA key-pair, then push the public key to the Secret Management Service. The latter, in turn, will provide it, at runtime, to the enclave running your Confidential Computing application.

To generate the key-pair, go to ~/iexec-projects and use the following SDK command:

Make sure your chain.json content is correct.

iexec result generate-encryption-keypair

This generates two files in .secrets/beneficiary/. Make sure to back up the private key in the file <0x-your-wallet-address>_key.

.secrets
├── beneficiary
│   ├── <0x-you-wallet-address>_key
│   └── <0x-you-wallet-address>_key.pub
...

Now, push the public key to the SMS:

iexec result push-encryption-key --tee-framework scone

And check it using:

iexec result check-encryption-key --tee-framework scone

Now to see that in action, you'd need to trigger a task and specify yourself as the beneficiary in the command:

iexec app run <0x-your-app-address> \
    --workerpool debug-v8-learn.main.pools.iexec.eth \
    --tag tee,scone \
    --encrypt-result \
    --watch

Wait for the task to be COMPLETED and download the result:

iexec task show <0x-your-task-id> --download

If you extract the obtained zip and try to read the content of the file iexec_out/result.zip.aes you will find it encrypted:

mkdir /tmp/trash && \
    unzip <0x-your-task-id>.zip -d /tmp/trash && \
    cat /tmp/trash/iexec_out/result.zip.aes

{% code title="iexec_out/result.zip" %}

)3�Xq��Yv���ȿ��zE�fRu<\�ݵm�m�����疞r�����c��(a������{{'���ܼ����͛�q/[{�����H��t>����������h��gD$g��\.�k��j�����"�s?"��h�J�_Q41�_[{���X�����������Ԛ��a���蘟v���E����r����肽
������Յ�]9W�TL�*����
          ���t��d����z��O`����!���e�&����snoL3�K6L9����%

{% endcode %}

Now you should decrypt the result by running:

iexec result decrypt <0x-your-task-id.zip>

A new zip file appears in the current folder under the name results.zip. Eventually, unzip it:

unzip results.zip -d my-decrypted-result

And you can see the content of your result file:

$ cat my-decrypted-result/result.txt
Hello, world!

Voilà! By finishing this part, you should be able to use confidential computing on iExec like a Ninja. All parts of the workflow are protected: the execution, the dataset, and the result.

You can go to the advanced section and learn more about managing orders on the iExec to effectively monetize your applications and datasets.