RAM (Relational Asynchronous Messaging)

Our patented breakthrough technology (RAM) uses a new paradigm that most people find surprisingly simple, intuitive, and natural. Relational asynchronous message processing treats the dynamic data flowing across an enterprise as time-stamped messages and uses familiar relational operators augmented to handle time windows to process those messages. By using time windows, RAM can aggregate, join, and process these potentially infinite message flows in meaningful time groupings, without having to wait for the last message to arrive.

This new and fundamentally better way of managing and processing dynamic enterprise data, while firmly grounded in the technology and theory of Relational Databases, extends that relational world to enable "querying the presently flowing and future data."

In this new paradigm, queries become active agents that execute continuously. They await the arrival of needed input data, processing it as soon as it arrives, and they generate outputs as soon as those outputs are ready. Everything is done on a continuous input-output basis. You can change these standing queries while they are running, causing immediate effects on their processing and their outputs.

RAMMS (RAM Management System)

SQLstream has developed the key software that enables you to create and manage RAM applications. We call this software system the RAM Management System (RAMMS).

The facilities and features that RAMMS provide for building and managing RAM applications to handle dynamic data are analogous to the facilities and features that RDBMS provide for handling relational data that are stored and managed within static tables.

Extracting the maximum value from your Dynamic Data Assets

Dynamic data are by definition constantly changing, flowing, renewed and replaced. It can be hard to manage and meet all of the necessary demands on systems handling dynamic data, to handle every transaction appropriately but at the same time enable meaningful analysis, reporting, and response.

SQLstream makes it possible, by providing you both the declarative SQL built to handle time windows and also the RAMMS support for efficiently processing and managing applications using it. RAMMS include support for the following features and facilities:

• Aggregating down and combining voluminous data streams into manageable and actionable compact data

• Continuously processing and querying all of your dynamic data flowing on the wire

• Executing complex SQL operations in parallel over specified time-windows

• Eliminating the latency of writing to disc and updating database indexes

• Automatically determining which data have changed and forwarding the results

• Generating alerts to subscribers

• Reliably delivering content-based results to subscribers

• Interfacing out-of-the-box to the widest set of data sources

Using SQLstream to gain these advantages is all the more attractive because of the economic benefits it affords.

Low cost ownership, development, and production

Production operations using applications built with SQLstream can run on low-cost commodity servers. The development of such applications leverages the broad base of highly skilled SQL designers and programmers. For SQL-familiar users, SQLstream requires little if any training, since SQLstream supports ANSI standard SQL/1999 and SQL/2003.

Instead of using complex programming to accomplish the integration of diverse systems, your teams can focus on high-level data management using declarative SQL, including the proven concept and power of SQL views.

With SQLstream and RAMMS, you rely on painless, proven technology in a self-optimizing data processing infrastructure that is also potentially self-healing in coping with system faults and outages. SQLstream was designed to support a system-wide ability to replan queries and remap them onto available hardware resources without suspending or corrupting execution.

Messaging Middleware, efficiency, and performance

SQLstream's RAMMS provides the next generation in Messaging Middleware by providing a high-level declarative messaging management system.

Rather than just operating at the low-level point-to-point and low-level topic-based Publish-Subscribe interface, SQLstream's RAMMS supports high-level auto-optimized relational operations. Whereas conventional Messaging Middleware operates on a message-by-message or record-by-record basis, SQLstream allows optimization of operations over potentially thousands of messages at a time. Not only is greater throughput achieved while maintaining low latency, you also gain the use of much higher level constructs and queries akin to those found in RDBMS such as Oracle™ and DB2™.

SQLstream also takes the strain out of managing time-sensitive operations and data by supporting a rich set of time-windowing operations. These facilities even allow time-windowed joins of message streams with joins to external database data, enabling the most efficient use of both statically stored and dynamically flowing data.

SQLstream provides previously unseen flexibility and performance by combining optimized binary representation coupled with automatic optimization over potentially thousands of messages.

Architecture

SQLstream's patented architecture was designed from the start to offer distributed declarative processing of relational asynchronous messages on commodity servers. The complexity of managing message processing across multiple distributed servers is entirely handled by SQLstream's RAMMS. It provides centralized management of all metadata and processing while allowing the actual RAM processing to execute on RAM Processors (RAMPs) located elsewhere on the network.

SQL Examples

SQLstream executes ANSI-standard SQL while incorporating extensions that support RAM streams. It is compliant with ANSI SQL:1999 and SQL:2003 standards, and supports ANSI-standard time windows. However, SQLstream's RAMMS executes its SQL over streaming relational asynchronous messages, effectively executing queries over future, flowing data rather than past, historical, stored data as queried by conventional RDBMS such as Oracle™ and DB2™. SQLstream provides a parallel, dataflow-based architecture where the data are pushed through the queries rather than vice versa as performed by RDBMS.

Adapters

To automate the generation of RAM data message streams as inputs to or outputs from RAMMS, SQLstream offers a rich set of adapters for developer use.

Drivers

SQLstream offers standards-compliant drivers for JDBC and ODBC, the industry standards for interactions with databases, such as inputting data into RDBMS. In the context of SQLstream, these drivers enable the insertion of messages into named RAM streams as opposed to the insertion of rows into named tables, as found in RDBMS. These drivers thus enable many existing applications to take immediate advantage of SQLstream's RAMMS with truly minimal application rewrite.

As the next generation of Messaging Oriented Middleware (MOM), SQLstream also offers a standards-compliant driver for JMS (Java Messaging Service), enabling messaging applications to be independent of any specific MOM platform.

Plug-Ins

SQlstream offers a rich set of dynamically loadable adapters, plug-ins with the following capabilities:

• Generate new input RAM data streams from external data sources,

• Generate new external out data streams from internal RAM streams, or

• Transform RAM streams using non-SQL programming modules and logic.

SQLstream's adapters include the following:

• Plain-text log files

• Log4J

• SOAP and Web-services

• HTTP

• RDBMS

• Email messages

• Text alerts

Software Development Kit (SDK)

SQLstream also offers an SDK whose flexibility, functionality, and expandability enables the easy construction of any required adapter.

Users can thus extend SQLstream's RAMMS to provide any kind of processing logic, enhancing and extending the already powerful SQL engine provided.