asfimport
commented
8 years ago Renaud Delbru (migrated from JIRA)
Discussion copied from the following dev thread
I would strongly recommend against "invent your own mode", and instead using standardized schemes/modes (e.g. XTS).
Separate from that, I don't understand the reasoning to do it at the codec level. seems quite a bit more messy and complicated than the alternatives, such as block device level (e.g. dm-crypt), or filesystem level (e.g. ext4 filesystem encryption), which have the advantage of the filesystem cache actually working.
@rmuir,
Yes, you are right. This approach is more complex than plain fs level encryption, but this enables more fine-grained control on what is encrypted. For example, it would not be possible to choose which field to encrypt or not. Also, with fs level encryption, all the data is encrypted regardless if it is sensitive or not. For example, in such a scenario, the full posting lists will be encrypted which is unnecessary, and you'll pay the cost of encrypting the posting lists. It is true that if the filesystem caches unencrypted pages, then with a warm cache you will likely get better performance. However, this also means that most of the index data will reside in memory in an unencrypted form. If the server is compromised, then this will make life easier for the attacker. You have also the (small) issue with the swap which can end up with a large portion of the index unencrypted. This can be solved by using an encrypted swap, but this means that the data is now encrypted using a unique key and not a per-user key. Also, this adds complexity in the management of the system. Highly sensitive installations can make the trade-off between performance and security. There are some applications for Solr that are not served by the other approaches.
This codec was developed in the context of a large multi-tenant architecture, where each user has its own index / collection. Each user has its own key, and can update his key at any time. While it seems it would be possible with ext4 to handle a per-user key (e.g., one key per directory), it makes the key and index management more complex (especially in SolrCloud). This is not adequate for some environments. Also, it does not allow the management of multiple key versions in one index. If the user changes his key, we have to re-encrypt the full directory which is not acceptable wrt performance for some environments.
The codec level encryption approach is more adequate for some environments than the fs level encryption approach. Also, it is to be noted that this codec does not affect the rest of Lucene/Solr. Users will be able to choose which approach is more adequate for their environment. This gives more options to Lucene/Solr users.
osnatShomrony
We would like to contribute a codec that enables the encryption of sensitive data in the index that has been developed as part of an engagement with a customer. We think that this could be of interest for the community.
Below is a description of the project.
Introduction
In comparison with approaches where all data is encrypted (e.g., file system encryption, index output / directory encryption), encryption at a codec level enables more fine-grained control on which block of data is encrypted. This is more efficient since less data has to be encrypted. This also gives more flexibility such as the ability to select which field to encrypt.
Some of the requirements for this project were:
What is supported ?
How it is implemented ?
Key Management
One index segment is encrypted with a single key version. An index can have multiple segments, each one encrypted using a different key version. The key version for a segment is stored in the segment info.
The provided codec is abstract, and a subclass is responsible in providing an implementation of the cipher factory. The cipher factory is responsible of the creation of a cipher instance based on a given key version.
Encryption Model
The encryption model is based on AES/CBC with padding. Initialisation vector (IV) is reused for performance reason, but only on a per format and per segment basis.
While IV reuse is usually considered a bad practice, the CBC mode is somehow resilient to IV reuse. The only "leak" of information that this could lead to is being able to know that two encrypted blocks of data starts with the same prefix. However, it is unlikely that two data blocks in an index segment will start with the same data:
Stored Fields Format: Each encrypted data block is a compressed block (\~4kb) of one or more documents. It is unlikely that two compressed blocks start with the same data prefix.
Term Vectors: Each encrypted data block is a compressed block (\~4kb) of terms and payloads from one or more documents. It is unlikely that two compressed blocks start with the same data prefix.
Term Dictionary Index: The term dictionary index is encoded and encrypted in one single data block.
Term Dictionary Data: Each data block of the term dictionary encodes a set of suffixes. It is unlikely to have two dictionary data blocks sharing the same prefix within the same segment.
DocValues: A DocValues file will be composed of multiple encrypted data blocks. It is unlikely to have two data blocks sharing the same prefix within the same segment (each one will encodes a list of values associated to a field).
To the best of our knowledge, this model should be safe. However, it would be good if someone with security expertise in the community could review and validate it.
Performance
We report here a performance benchmark we did on an early prototype based on Lucene 4.x. The benchmark was performed on the Wikipedia dataset where all the fields (id, title, body, date) were encrypted. Only the block tree terms and compressed stored fields format were tested at that time.
Indexing
The indexing throughput slightly decreased and is roughly 15% less than with the base Lucene.
The merge time slightly increased by 35%.
There was no significant difference in term of index size.
Query Throughput
With respect to query throughput, we observed no significant impact on the following queries: Term query, boolean query, phrase query, numeric range query.
We observed the following performance impact for queries that needs to scan a larger portion of the term dictionary:
We can see that the decrease of performance is relative to the size of the dictionary scan.
Document Retrieval
We observed a decrease of performance that is relative to the size of the set of documents to be retrieved:
Known Limitations
compressed stored field do not keep order of fields since non-encrypted and encrypted fields are stored in separated blocks.
the current implementation of the cipher factory does not enforce the use of AES/CBC. We are planning to add this to the final version of the patch.
the current implementation does not change the IV per segment. We are planning to add this to the final version of the patch.
the current implementation of compressed stored fields decrypts a full compressed block even if a small portion is decompressed (high impact when storing very small documents). We are planning to add this optimisation to the final version of the patch. The overall document retrieval performance might increase with this optimisation.
The codec has been implemented as a contrib. Given that most of the classes were final, we had to copy most of the original code from the extended formats. At a later stage, we could think of opening some of these classes to extend them properly in order to reduce code duplication and simplify code maintenance.
Migrated from LUCENE-6966 by Renaud Delbru, 12 votes, updated May 22 2020 Attachments: Encryption Codec Documentation.pdf, LUCENE-6966-1.patch, LUCENE-6966-2.patch, LUCENE-6966-2-docvalues.patch Linked issues:
10419
3304