Cloaked Search

A search index is much like a book index. It contains a list of words that point back to the locations where they appeared. But this data can be used to recreate the entire book. And in most cases, the search service actually retains the original book, too. Yet for most companies, the data in the search index isn’t protected at the same level as other data stores.

Encrypted search is a technique that encrypts both the raw data and the index. The data can still be searched over without decrypting it, but anyone who wants to search it must have the right key. The server is blind to the data it holds and to the meaning of the searches that are made.

You choose which fields to encrypt on which indices. You can do everything on unencrypted fields, and nearly everything on the encrypted ones.

Encryption is best suited to text fields such as name, email, address, health data, and user-generated content. Nearly all of the functionality you love from Elasticsearch and OpenSearch can be used over the encrypted data.

While not everything is supported, for example, you can’t do regular expression searches over the encrypted data, all of the most commonly used features are.

And most of the time, you don’t even have to change your code.

Single key standalone mode

In standalone mode, a single key is loaded into the proxy at startup and that key is the root of trust for securing all the data. This is the simplest deployment available. Simply configure which fields you want encrypted in which indices, then point your code to the proxy instead of the search service.

Multi-key advanced key management

When you need different keys for different segments of data or when you want to make sure the service provider (eg, Amazon, Google) can’t gain access, then you can integrate SaaS Shield as a key orchestration component for keeping master keys in one or more independent key management servers. This is especially useful for multi-tenant SaaS and customer-held encryption key patterns.