UNICOM PIE/CICS-NetMizer

 

PIE®/CICS-NetMizer

Customized Compression of 3270-Based Data Streams

 

OVERVIEW

PIE/CICS-NetMizer is an optional PIE/CICS component that reduces the size of 3270 data streams before they are sent across a network. PIE/CICS-NetMizer uses a combination of data compression techniques and optimization methods to reduce the overall volume of network traffic. As a result, network congestion is reduced and the average transaction response time improves at most sites with less inherent delay. PIE/CICS-NetMizer uses a combination of compression and optimization algorithms. Compressed data streams can be as much as 60% to 95% smaller than their original size. PIE/CICS-NetMizer also uses a variety of optimization methods that reduce the number of data streams required to sustain current screen images on user terminals.

Inbound Suppression Optimization formatted screens send only modified fields from the terminal to an application.

A modified field is indicated by two methods:

• User changes the field at the terminal
• Application marks the field as modified

Applications mark modified fields to ensure they are returned from the terminal, whether the field is changed by a user or not. Usually, this is done to avoid the coding required to keep the fields in a commarea. There is a significant amount of network overhead to send unchanged modified fields back and forth between an application and a terminal.

PIE/CICS-NetMizer reduces the overhead of sending an inbound data stream from the terminal to an application. On the outbound data stream from the application to the terminal, PIE/CICS-NetMizer records the fields the application marked as modified. It then resets the fields to nonmodified and sends the data stream to the terminal. Users perform their processing and press an AID key to return the data stream. As always, only changed fields are returned over the network. PIE/CICS-NetMizer then intercepts the data stream. If any of the fields the application marked as modified are missing from the data stream, they are added back to the data stream and returned to the application.

Inbound Suppression Optimization results in the following:

• All data is sent to the terminal.
• Only changed fields are received back from the terminal.
• All marked fields are returned to the application.

The application receives all requested fields, but the network sends only the modified data.

Read Buffer Suppression

Some applications issue READ BUFFER requests to the terminal. In response, the terminal sends a copy of the buffer’s contents back to the application. For 3270 model 2 terminals, this is always 1,920 characters. A copy of the entire buffer’s contents is sent, including all blank and null characters. PIE/CICS-NetMizer can intercept READ BUFFER requests and send a copy of the terminal buffer stored in memory. All read buffer requests by applications are suppressed, reducing the amount of network traffic.

Implicit Benefits of Data Stream Optimization

The previous types of optimization techniques have the net effect of reducing the size and number of data streams sent over a network. In addition, there are other implicit benefits to overall system performance by reducing the size and number of data streams.

Reduced RU chaining

VTAM network data streams are transmitted in packets known as Request Units or RUs. Each RU contains VTAM control information as well as the data being sent to or from a device. RU sizes vary, depending on the network configuration. For example, an installation might specify RU sizes of 256 bytes, while another installation specifies RU size of 1024. Thus, when an application or device prepares to send data across a network, it must first divide the data into multiple RU packets and chain them together. This chaining process requires some CPU processing. Also, sending chained packets is slower than the total time to send all individual packets. Sending smaller data streams across a network reduces the amount of required of RU chaining and consequent CPU processing.

Reduced VTAM data buffering

Data stored within an application’s address space must be sent to VTAM before being distributed by a network. The longer the message, the more buffering is required. The shorter the message the less buffering will be required.

 

AT-A-GLANCE

• Repetitive character optimization
• Imaging optimization
• Semantic imaging
• Inbound suppression optimization

 

FEATURES

Compression Levels

• Terminal optimization
• Session and menu optimization
• Omitting data stream optimization
• Application optimization