Memory and Disk Capacity Estimation Guidelines

Administrator
This manual is in pilot operation.

fusion_place retains ledger data in memory. For smooth system operation, it is necessary to prepare sufficient memory capacity according to the number of ledger data items. This section explains how to estimate the required memory capacity and the necessary disk capacity.

The memory and disk requirements estimated in these guidelines are intended as rough estimates. While they are intended to assist in decision-making, actual values may differ significantly depending on system configuration, application design, and user access levels. Fusion Inc. does not guarantee that these estimates will be appropriate for your environment using fusion_place. Please verify memory and disk consumption by conducting operational verification with actual data on the actual system.

For methods to check memory consumption and adjust the amount of memory allocated to fusion_place, please refer to the Memory Management section at Memory Management.

Memory Capacity

Ensure that the total amount of unused physical memory exceeds the following amounts when neither the fusion_place server nor client is running (although on standalone machines or client machines, Microsoft Excel should be running).

Items Consuming Capacity

Standalone-based

Web Server-based

Standalone Machine

Server Machine

Client Machine

Program (including JRE/JDK)

About 120 MB

About 80M MB

About 120M MB

Server Data (The amount of memory required to continuously retain ledger data)

See below "Method for Estimating Ledger Data Capacity"

Server Work Data (The amount of memory required to retain temporary work data)

Depends on the number of users requesting processing simultaneously (for standalone, 1) and the number of data items processed at once. Estimate the percentage of server data obtained or displayed at once, multiply that percentage by the above server data capacity, and then multiply by the number of users.

Client Work Data (The amount of memory required to temporarily retain data of documents to be displayed/printed)

Depends on the number of data items processed at once and the form design. Estimate the percentage of server data obtained or displayed at once, multiply that percentage by the above server data capacity, and then multiply by a safety factor of about 2.

(Same as Standalone-based)

Disk Capacity

Items Consuming Capacity

Standalone-based

Web Server-based

Standalone Machine

Server Machine

Client Machine

Java (JRE or JDK)

JRE

About 80 MB

JDK

About 160 MB

Program Files

About 40 MB

About 15 M

Database

See below "Method for Estimating Ledger Data Capacity"

Others

Several MB
(Local copies of master data, etc.)

Method for Estimating Ledger Data Capacity

When the number of ledger data items is large, most of the "Server Data" in memory capacity and the "Database" in disk capacity are the capacity of ledger data. Estimate the ledger data capacity as follows.

Estimating the Number of Ledger Data Items

The number of ledger data items is the number of combinations of the lowest-level members where data actually exists in the ledger. However, note the following points:

  • Ignore relative periods and display format dimensions, and count the combinations of lowest-level members for other dimensions.

  • For system-reserved members of the change account dimension, count only Beginning Balance/#BEG, Cumulative Balance/#CBAL, and Ending Balance/#END (do not count other system-reserved members).

  • Count each ledger separately and aggregate the results.

  • For each ledger edition, estimate the number of data items changed since creation from the parent ledger edition, and sum these totals (each ledger edition retains differential data from the parent ledger edition).

Estimating the Memory Capacity of Ledger Data Based on the Number of Data Items

For amount/numeric data, if there are 100,000 data items, the memory capacity of ledger data is about 50M bytes (500 bytes/item). [1][2]

If the number of data items exceeds 100,000, the data capacity increases almost proportionally to the number of items.

Conversely, if the number of items is smaller, the capacity decreases, but more than proportionally. For example, for memory capacity, if the number of data items is 10,000, the required capacity is not 5 MB (1/10 of 50 MB for 100,000 items), but about 10 MB.

These estimates assume a minimum period unit of 12 per fiscal year. If the minimum period unit exceeds 12 (e.g., including quarterly adjustment periods to 16), the data capacity increases almost proportionally to the minimum period unit.

Estimating Database Capacity

Ledger data is retained in the database, including change history, so the database capacity of ledger data is larger than the above memory capacity. For example, if an average of 1 change is made in the workspace and history is deleted after publication, the database capacity of ledger data is about 3 times the memory capacity (Ledger Edition History + Workspace History x 2). Besides this, the database also includes master data, etc., but this is usually several tens of MB.


1. If ledger data contains a large number of string values, the data capacity tends to be larger depending on the length of the strings.
2. This estimate is based on using Oracle Java 32-bit Java. If using 64-bit Java or other Java, the required memory capacity may increase.