Attribute

Size

Number of records in a table

274,877,906,944

Record size

• Paged Data: 64K
• In-Memory: Unlimited
• Memory-Persisted: Unlimited

Number of tables per data application

2048

File size

• 4 GB file or limited by the operating system file size constraints. 
• 4.2 TB using Page Service with striped file Page Adapter

Memory footprint for Runtime Services

Min: 24KB
Max: Depends on application since it is converted to C code

Number of fields in a table

2 billion

Number of indexes per table

Unlimited

Number of bytes per index key

• BTree: 64K
• In-Memory: Unlimited

Number of tables referenced per transaction

No limit

Attribute

Characteristics

Language

SQL and Procedural SQL (similar to PL/SQL), with embedded extensions, including:

• Manage streams of data
• Table Types
• Control Structures (IF, THEN, ELSE, LOOP, WHILE, EXIT, …)
• Exporting DeviceSQL function name to C
• Importing of C Code as DeviceSQL function
• Choose storage location per table

Table JOIN support

Yes (INNER)

Cursor Support

Internal and manual control (Open, Fetch, Close)

Exception Handling

• Handle exceptions directly in functions/procedures
• Create and raise user-defined exceptions
• Pass error codes to C application code

International Language Support

Roman characters (ASCII) and Unicode (UTF-16)

APIs for Application calls 

Developer-defined API, Dynamic C API, Interpreted SQL API

Critical Section Support

Important code sections can be protected within DeviceSQL code or from within C application code

Data Types

• Simple Types: BOOLEAN, VARCHAR, NVARCHAR, SHORT, CSHORT, INT8, INT16, INT32, LONG, BIT8, BIT16, BIT32, DOUBLE, CDOUBLE, DATE, BINARY, TIMESTAMP, FLOAT64, PHONECHAR, POINTER, VECTOR_INT8, VECTOR_INT16, VECTOR_INT32, VECTOR_BIT8, VECTOR_INT16, VECTOR_INT32, VECTOR_DOUBLE,
• Complex types:  RECORD, TABLE, STREAM, AGGREGATE

Event Hooks

• Available to run C functions and procedures
• Control maximum memory usage
• Interrupt long running queries

Functionality

Values

Host Development Platforms

Windows, Linux, Solaris

Target CPU

Supports virtually any target CPU, including multicore CPUs, Commonly supported CPUs include:

• ARM, PowerPC, MIPS, SuperH, x86 and others

Target OS

DeviceSQL Runtime Services available for any target OS including Linux, VxWorks, Windows, u-ITron, QNX and others

Runtime Library Type

Static Library (dynamic on demand)

IDEs Supported

• Microsoft® Visual Studio™
• GNU development toolchain and command-line operation
• Wind River® Workbench
• ARM® RealView® Development Suite
• Renesas HEW
• QNX® Momentics® Development Suite
• Freescale® CodeWarrior™

Transaction Support

• COMMIT, ROLLBACK, and AUTOCOMMIT
• ACID compliance
• Create Read-Only or Read-Write Transactions

Compression

Compresses paged data. Algorithms available are:

• ZLIB
• BZ2

Additional compression algorithms and methods are available from Encirq professional services

Communication

Data Streams layer communication support for TCP sockets, FIFO and shared memory

Runtime Backup and Restore

• Available, using SM_COPY, SM_OPEN, SM_CLOSE
• Defragmentation performed during backup

Storage Management

• Services-based (only use what is required)
• Store fixed and variable sized data
• Write to any persistent storage type, such as  ROM, NVRAM, Flash, Disk, EEPROM
• Table data can reside in-memory, memory-persisted, and paged-in as needed
• Customizable abstraction layer
• Support permanent and temporary tables
• Multiple persistent file support
• Storage can be managed per table
• Each manager can be open/closed as needed
• Configurable tuning parameters, such as stripe size

Index Management

• Services-based (only use what is required)
• Several different index types (Hash, BTree, AVL)
• Reside in memory or paged-in as needed
• Create indexes using user-defined algorithms
• Configurable tuning parameters, such as hash bucket size
• Open and close indexes as required
• Create and drop using SQL syntax

Memory Management

• Memory Management service allows user-defined functions for controlling memory allocation and de-allocation
• Manage data entirely in memory (In-Memory)
• Load persistent data into memory at database open (Memory-Persisted)
• Load data into memory as needed (Paged)
• Parameters for constraining memory usage
• Configurable tuning parameters, such as buffer and page size

Compaction Algorithms

For data types, storage managers, and indexes

Memory Optimizations

For SELECT query results, indexes, and paging of data

Thread-safe Support

Native thread-safe or coded multiple thread access to shared data

Multi-Process Support

• Multiple concurrent readers can access shared data from same database instance
• Processes can access separate database instances