Overview

This section explains the basic concepts and requirements of Workgroup engines. If you need more in-depth information about the Workgroup engine, refer to the Advanced Operations Guide. The Advanced Operations Guide contains detailed technical information about the Workgroup engine, setting up a Gateway configuration, and re-directing locator files.

Every computer that may be used to access the same data at the same time must have a Workgroup engine installed on it.

Only database server engines can enforce OS level file security based on the privileges assigned to the login user name. The Workgroup engine does not attempt to do this. In a small office, where Workgroup engines are most common, this can be considered a plus because they are usually short on networking experts, and the fewer barriers to successful data access the better.

There are three main configurations in which you would want to use the Workgroup engine.

The first configuration takes place when all the data is located on a single computer with a Workgroup engine installed, and there is limited sharing of data. This configuration is roughly equivalent to a small client/server configuration.

Another situation when you would want to use the Workgroup engine is when the data is distributed among the workstations. This is called a peer-to-peer topology. This configuration is used when each application typically stores much of its own data on the local hard drive, but periodically needs to access data from other workstations or share its own data with others.

In this configuration, each computer shares its data directory or directories. Any computer that needs access to that data maps one or more drives to the shared data directories. Then the Workgroup engine on each computer acts as a mini-server engine to read/write all changes to the data files on that machine.

The third topology requiring the use of the Workgroup engine is when the data is stored on a file server where there is no MicroKernel engine. This can be a Unix server or other type of network file server that gets backed up regularly, but cannot support a MicroKernel engine. In this situation, the first Workgroup engine that opens files in a directory on the server becomes the Gateway to each file in that directory. The other workstations access the data in a client-server fashion through that Gateway engine.

The Gateway engine for a given directory identifies itself by creating a file named ~PVSW~.LOC in that directory. This file is called a Gateway locator file and contains the network name of the computer where the Gateway engine is located. Other Workgroup engines attempting to access this data read the locator file to find the name of the engine they must communicate with in order to access the data.

You can ensure that the same engine always services the files in a given directory by making the locator file read-only. This is called a static gateway, also referred to as a fixed gateway. See To Set up a Fixed Gateway for more information.

The Gateway engine acts as a server engine as it reads and writes pages to the data files, allowing it to make the most use out of its cache. The Gateway feature is designed so that the ownership of any particular directory can change whenever the current gateway engine has no more client applications with any files open in that directory. When the last data file is closed in a directory by a given database engine, the engine releases and deletes the locator file. When the next engine opens a data file, that engine becomes the new gateway to the directory where the data file(s) resides.

A Gateway engine is a Workgroup engine that acts as the sole point of access to all data files in a particular directory on a remote file server. If several Workgroup engines are accessing the same database at the same time, they do not all open the files simultaneously, nor do they share the files. Rather, the first Workgroup engine to access that database becomes the temporary “owner” of those files, and all other Workgroup engines must access the data by contacting the Gateway engine. Only the Gateway engine has the files open and reads/writes the files. The other Workgroup engines act as clients, making requests to the Gateway engine acting as a mini-server engine.