tlSpeed is an extremely fast database engine designed to meet unique demands of medium and large customers.

tlSpeed is unique as it uses distinctly large index structures enabling almost immediate access to the searched data.

tlSpeed indices are created by using very effective, RAM-oriented algorithms so as to avoid accessing slow media such as HDDs.

The tlSpeed disk repository can be dynamically extended as the need arises. It is always possible to add a new storage device (a single drive or a whole disk array) to the storage area. This allows database centres to be built in a more flexible and less time consuming way. tlSpeed makes use of binary data compression, reducing required storage area twofold.

Data integrity and fault tolerance result from a special construction of the database file system. Loss of a fragment of the database does not influence its overall performance. Restoring the deleted fragment is easy and fast - either a backup or source files may be used to achieve fault-recovery without any down-time for the engine.

Safety of the system based on the tlSpeed database is ensured by the following features.
Firstly, the integrity of the incoming data stream can be verified in real-time by using CDR sequence numbers. Secondly, the access to search API is protected by a security filter which requires strong authorization of the accessing party and provides optional data encryption. Moreover, the database may be configured to protect against bulk-quantity data extraction in order to limit possible damages by a rogue employee.

Backup procedures - the structure of the tlSpeed database allows to make a full backup on-line, without any down-time. In order to perform an incremental backup it is enough to copy new files in the database file system at any chosen moment.

The tlSpeed database is designed for:

• telecommunications companies,
• MVNOs,
• ISPs,
• large and medium e-commerce,
• large Web portals,
• Internet-based payment and/or micropayment brokers,
• other companies which process, archive and make available huge amounts of data incoming in real-time.

Typical deployment scenarios of the tlSpeed database include the following requirements:

• each data record has its timestamp (CDR, financial transaction log, user behaviour log);
• records with similar timestamps are usually fed into the system together (although importing past data is possible, even in large batches);
• incoming data flow is massive (measured in millions of records per hour).