Airlock IAM Performance Tuning

There are various aspects that influence performance and ability to scale of an Airlock IAM system depending on the precise usage scenario. This article highlights the most common bottlenecks.

Audit Log

Airlock IAM up to 6.4.x digitally signs audit log messages by default. As of IAM 7.0 audit log signing is disabled by default. The signature ensures that manipulations of log messages are detectable. The generation of log sequence numbers and digital signatures limits scaling with the number of CPU cores. Disabling audit log signing can speed up the whole system significantly. Audit log messages are still written - but not signed anymore. Starting with Airlock IAM 7.0 the signing of audit log messages is by default disabled.

---

Database Selection

The system requirements page lists the supported databases. Some of them are known to scale well with the number of users or tokens. Others are fine for smaller scenarios only.

The supported H2 database has limitations:

• H2 uses only one index per logical table at a time in each query.
• The H2 cluster feature imposes additional limitations on performance (result sets are completely loaded into memory).

---

Database Monitoring

 

---

Database Indices

To scale well, a database needs indices that are based on the most frequent use cases. Since Airlock IAM is highly configurable regarding database fields, the indices have to be set with the knowledge of the field configuration and usage.

Airlock IAM versions 6.4, 6.4.1 and 7.0 define additional indices and primary keys in the database schema setup scripts. The database schemas are not updated automatically when updating the Airlock IAM software. Changes to the database schemas have to be performed manually. Recommended changes for all databases are documented in the IAM manual: Database Schema History.

Recommended changes for all databases for all versions of Airlock IAM - even for versions prior to 6.4:

CREATE INDEX ta_index ON token_assignment (ta_user, ta_token_id);
CREATE INDEX serial_id_index ON token (serial_id);
CREATE INDEX token_type_index ON token (type);
CREATE INDEX obsoletes_token_id_idx ON token (obsoletes_token_id);

Additional changes for MySQL/MariaDB:

CREATE INDEX generic_data_element_1_idx ON token (generic_data_element_1(40));
CREATE INDEX generic_data_element_2_idx ON token (generic_data_element_2(40));
CREATE INDEX generic_data_element_3_idx ON token (generic_data_element_3(40));
CREATE INDEX generic_data_element_4_idx ON token (generic_data_element_4(40));
CREATE INDEX generic_data_element_6_idx ON token (generic_data_element_6(40));
CREATE INDEX generic_data_element_7_idx ON token (generic_data_element_7(40));

Additional changes for Oracle/H2:

CREATE INDEX generic_data_element_1_idx ON token (generic_data_element_1);
CREATE INDEX generic_data_element_2_idx ON token (generic_data_element_2);
CREATE INDEX generic_data_element_3_idx ON token (generic_data_element_3);
CREATE INDEX generic_data_element_4_idx ON token (generic_data_element_4);
CREATE INDEX generic_data_element_6_idx ON token (generic_data_element_6);
CREATE INDEX generic_data_element_7_idx ON token (generic_data_element_7);

The indices on the generic_data_element_XX columns cannot be created with the default schema for MS SQL Server databases, because NVARCHAR(MAX) columns can't be specified as key columns for an index: MS SQL documentation.

Recommended changes for all databases for Airlock IAM versions starting with 6.4:

ALTER TABLE token_assignment ADD CONSTRAINT fk_ta_token_id
      FOREIGN KEY (ta_token_id) REFERENCES token (token_id);
ALTER TABLE token ADD CONSTRAINT fk_obsoletes_token_id
      FOREIGN KEY (obsoletes_token_id) REFERENCES token (token_id);
ALTER TABLE token ADD CONSTRAINT fk_activates_token_id
      FOREIGN KEY (activates_token_id) REFERENCES token (token_id);

---

DB Statistics

DB systems use optimziers to tune queries on their own. The decisions of the query optimzier base on database usage statistics. The statistics have to be up-to-date to result in the best decisions.

---

DB Connection Pool

Connections from Airlock IAM to the DB are handled by connection pools.

The idea is to run only as many parallel requests in the DB as it can handle at a time:

• Using too few DB connections will result in a bottleneck (wait events in IAM).
• Using too many DB connections can result in DB wait events and lower the overall performance as well.

Each configured DB connection has its own pool size. A separate connection pool is used by each module (Loginapp, Adminapp, SOAP Interface, Service Container, Transaction Approval) and connection pool configuration (e.g. pool size). The Loginapp typically needs more DB connections than the Adminapp.

See https://github.com/brettwooldridge/HikariCP/wiki/About-Pool-Sizing for more information about connection pool sizing.

---

Current JDBC Driver

Database vendors provide JDBC drivers for specific versions of their databases. JDBC drivers are not bundled with Airlock IAM and must be downloaded and deployed separately. It is essential to use the  latest JDBC driver that matches the database version currently in use. Unexpected behavior including degraded performance can result from an inappropriate driver version.

---

Hash Algorithms

Some hash algorithms, such as "scrypt" are expensive by design in order to make brute-force attacks on hash databases unfeasible, i.e. increase the security of the hashed data. On the other hand, choosing an expensive hash algorithm limits the number of request that a system is able to process per time unit. Other hash algorithms, such as "sha256" are pretty fast and therefor easier to bruteforce.

Airlock IAM uses hashes for various use cases. Depending on the use-case a different hash algorithm is suitable.

See selection and parametrization of hash functions for further information and tuning options.

---

Current Version

New Airlock IAM versions are often bringing additional performance optimizations. Check for the current version. Especially versions 6.3.6 and 6.4.1 are containing valuable improvements for REST calls and token persistence.

---

Example of Dimensioning and Scaling

Assumed Hardware:

• Server with 2 CPUs each with 8 cores at 2.5 GHz.

Assumed Configuration:

• password with scrypt hashes as the first factor
• MTAN as the second factor
• Audit Log signing is switched off
• the DB connection pool has a size of 16 and fits well the performance of the DB server
• the "general indices" mentioned above are created.

Expected Performance:

• 50-90 entire login/logout roundtrips per second

Airlock IAM scales well with the number of CPUs and cores. However CPU threads won't help, since the system is CPU-bound (scrypt). CPU threads are beneficial if a CPU core has to wait for I/O.

For further scaling it is possible to loadbalance multiple Airlock IAM behind Airlock WAF.

 

---

NUMA (Non Uniform Memory  Access) Performance Tuning Option

Non Uniform Memory Access is a result of modern hardware architectures. If configured incorrectly, it may lead to inefficient memory access behaviour for processes. Whether or not a system may profit from enabling NUMA support depends on the actual setup of hardware, operating system, virtualization and container technology. It is recommended that during performance tuning the impact of enabling NUMA support is measured directly on the target system.

• To verify, if the underlying platform uses NUMA, run this command: dmesg | grep -i numa
• To enable NUMA support add the option "-XX:UseNUMA" in "iam.java.opts" in instance.properties