GIS operations rate among the heaviest data movers across networks (joined by document management and video conferencing enterprise solutions). Geographical information systems are data heavy because geography is data rich, and the beauty of GIS analysis is in its ability to examine high volumes of data quickly and turn it into information that is useful to you (an information product). A variety of factors affect performance of GIS enterprise solutions across the network.
Performance Factors
Key factors affecting network load to consider for various application architectures are listed in the below table:
| Application Architecture | Network Load | Mbit Per Display | Comments |
|---|---|---|---|
| Rich Client Applications | Heavy | 10 | Typically only recommended for LAN due to chattiness |
| Rich Client Applications via Citrix (Image) | Medium | 5 | Citrix or Terminal service connections allow for a robust user interface while reducing network load of WAN |
| Rich Client Applications via Citrix (Vector) | Light | 3 | Citrix or Terminal service connections allow for a robust user interface while reducing network load of WAN |
| Web Applications (Dynamic) | Medium | 2 | ADF based applications typically have a heavier network load then REST API applications |
| Web Applications (Cached) | Medium | 1 | ADF based applications typically have a heavier network load then REST API applications |
| Mobile Applications | Light | 0.05 |
Only incremental updates are periodically sent |
| Services (REST) | Medium | 2 |
REST based services can be designed for lightweight loads |
| Services (SOAP) | Medium | 3 | SOAP is a more robust mechanism for communication and therefore has a heavier footprint on the WAN |
Tuning and Optimization
Significant factors that you can adjust that affect network load are:
- Internet browser caching policy - Some organizations may disable local caching, increasing network load
- Percent of new users - New users have no data cached on their local systems, increasing network load
- Image Size - The smaller the image displayed on the client, the less network traffic is required
- Image Format- JPEG is typically lightest for raster and PNG lightest for vector
- Blending Location - Applications can be designed to blend images at client or server-side. Client-side blending increases the network loads at the clients (typically most constrained), server-side blending increases the network loads at the server/data center.
LAN Performance
Since GIS applications typically transfer large amounts of traffic across the network, the following are recommended throughput capabilities for LAN components:
- Clients - 100 Mbit/sec
- Servers - 1 Gbit/sec
WAN Performance
Additional considerations are necessary for Wide Area Networks (WAN), due to their typically smaller bandwidth and higher latency. A key performance consideration for enterprise GIS solutions across WANs is the total amount of data that can be transmitted over the network in a given amount of time. This quantity is determined by a combination of bandwidth and latency. Bandwidth is the speed of transmission over a network connection in kilobits per second. Latency refers to the amount of time it takes in milliseconds to transfer data from one point to another. Both of these factors combine to determine the amount of data that can be transmitted in a certain amount of time over the network. The product of these two factors directly affects user perception of how long it takes to process a transaction.
When evaluating your network connections, you need to evaluate bandwidth and latency, recognizing that although some types of network connections can maximize bandwidth, they may increase latency. For example, a satellite connection may offer high bandwidth, but latency may suffer when compared to ground connections such as frame relay or dial-up Integrated Services Digital Network (ISDN).
Minimizing the quantity of transactions performed with a client/server app can help significantly. Often, though, it is impractical for an IT shop to change the way a client/server app is written. Citrix and Terminal server can help with this, because the numerous transactions can be performed across a slow link. ICA or RDP traffic (the protocols that Citrix and Terminal server use) are relatively well behaved and work better across slow links that are prone to latency issues. Many large Citrix implementations have been successfully performed with Esri products. The most common reason for implementing a Citrix solution is to improve application performance across WAN connections with minimal bandwidth and/or high latency.
Some customers have utilized components like NetScaler or BranchRepeater (formerly WANScaler) to improve their throughput and others exposing large scale solutions on the Internet, utilize acceleration service providers such as Akamai, which can distribute cached information from your product across datacenter scattered throughout the world. It is not unusual to see reductions of not only network load, but also server loads with these solutions in the range of 30%. The effectiveness of a particular solution can vary significantly depending on your specific application, so ensure testing and validation are performed before assuming improvements.