Home
Tuning Guide
An installation of Unica Interact consists of several components including third-party tools (such as web application servers, databases, and load balancers) and components such as Unica Platform and Unica Campaign. All of these components have several properties, features, and settings you can configure to improve performance.
About tuning Unica Interact for best performance
An installation of Unica Interact consists of several components including third-party tools (such as web application servers, databases, and load balancers) and components such as Unica Platform and Unica Campaign. All of these components have several properties, features, and settings you can configure to improve performance.
Cache management
In a large-volume Unica Interact environment where you are using a large number of runtime servers, you can use cache management software to share the runtime load among the servers and improve the real-time performance of the runtime server group as a whole. When you enable cache management solution comes with Unica Interact you can use multiple run time servers fronted by a load balancer. The load balancer balances the workload across the runtime servers in the group, and helps to maintain something called session affinity, which means that when an incoming session is handled by runtime server A, then when the same user issues additional requests, those requests are fulfilled by the session on server A. This method of balancing connections among the runtime servers in the group can improve performance considerably, but has some limitations because all of the session information is maintained in memory, and the memory limits of each Java™ virtual machine (JVM) restrict how much session information can be maintained. Unica Interact supports two cache management solutions by default:
Working with Ignite
To improve performance for your Unica Interact runtime server group, you can configure Ignite by modifying a series of configuration properties in Unica Platform.
Supporting a greater number of concurrent sessions in Ignite
When you are using Ignite as the cache manager in some Unica Interact environments, a high number of concurrent sessions may cause the Unica Interact runtime to exceed its available memory, causing a system slowdown or out-of-memory error. An out-of-memory situation is more likely if you have increased the maxEntriesInCache configuration parameter (Unica Interact > cacheManagement > Caches > InteractCache > maxEntriesInCache) to a number higher than the default setting, although it can occur even with the default setting of a maximum of 100,000 sessions. To avoid this issue, you can reduce the maxEntriesInCache value, or you can follow the instructions here to modify the system memory cache to roll the cached memory over to disk storage. This modification allows many more concurrent sessions than would otherwise be possible.

Tuning Guide
An installation of Unica Interact consists of several components including third-party tools (such as web application servers, databases, and load balancers) and components such as Unica Platform and Unica Campaign. All of these components have several properties, features, and settings you can configure to improve performance.
- About tuning Unica Interact for best performance
  An installation of Unica Interact consists of several components including third-party tools (such as web application servers, databases, and load balancers) and components such as Unica Platform and Unica Campaign. All of these components have several properties, features, and settings you can configure to improve performance.
  - Cache management
    In a large-volume Unica Interact environment where you are using a large number of runtime servers, you can use cache management software to share the runtime load among the servers and improve the real-time performance of the runtime server group as a whole. When you enable cache management solution comes with Unica Interact you can use multiple run time servers fronted by a load balancer. The load balancer balances the workload across the runtime servers in the group, and helps to maintain something called session affinity, which means that when an incoming session is handled by runtime server A, then when the same user issues additional requests, those requests are fulfilled by the session on server A. This method of balancing connections among the runtime servers in the group can improve performance considerably, but has some limitations because all of the session information is maintained in memory, and the memory limits of each Java™ virtual machine (JVM) restrict how much session information can be maintained. Unica Interact supports two cache management solutions by default:
    - Working with Ehcache
      To improve performance for your Unica Interact runtime server group, you can configure Ehcache by modifying a series of configuration properties in Unica Platform.
    - Working with Ignite
      To improve performance for your Unica Interact runtime server group, you can configure Ignite by modifying a series of configuration properties in Unica Platform.
      - Modifying Ignite Configuration Settings
        When you specify that an Unica Interact runtime server should use the built-in cache manager called Ignite to improve performance, you can configure the settings used by Ignite to optimize its value to your runtime server group.
      - Supporting a greater number of concurrent sessions in Ignite
        When you are using Ignite as the cache manager in some Unica Interact environments, a high number of concurrent sessions may cause the Unica Interact runtime to exceed its available memory, causing a system slowdown or out-of-memory error. An out-of-memory situation is more likely if you have increased the maxEntriesInCache configuration parameter (Unica Interact > cacheManagement > Caches > InteractCache > maxEntriesInCache) to a number higher than the default setting, although it can occur even with the default setting of a maximum of 100,000 sessions. To avoid this issue, you can reduce the maxEntriesInCache value, or you can follow the instructions here to modify the system memory cache to roll the cached memory over to disk storage. This modification allows many more concurrent sessions than would otherwise be possible.
      - Failover Ignite
        When distributed Ignite is used, the data is saved across the whole grid. If the instance hosting a session is stopped, that session data will be lost and all subsequent requests for that session will fail. To achieve failover, you need to set parameter numberOfBackups under Affinium|Interact|cacheManagement|Cache Managers|Ignite|Parameter Data. Its value is the number of backup copies of the data in the cache. A higher value will have better failover protection and better read performance with the cost of lower write performance. Therefore, it should be chosen based on the use cases.
  - Interact API
    Use the Java™ Serialization API instead of SOAP API. The Serialization API provides better throughput (can be 5-10 times more depending on the application configuration) and a shorter response time.
  - Installation and network configuration
    If the Unica Interact server is required to maintain session data across multiple Unica Interact API calls and you are using the Ehcache cache manager to improve performance, use sticky load balancing and local session management instead of distributed session management.
  - Interactive flowchart management
    Every interactive flowchart requires at least one thread to run. You can monitor a live system to see whether there are enough threads for all the interactive flowcharts.
  - Service tuning
  - Web application server tuning
    Within Unica Interact, you tune the web application by modifying JVM arguments and connections. The JVM arguments affect throughput and startup time. The number of connections you use is determined by the features you have enabled.
  - Tuning terminology
    System tuning has specific definitions for common terms.
  - Database tuning
    Database tuning involves adding indexes to specific tables and updating statistics.
  - ETL Tuning
    When you configure the contact and response history module, the module uses a background Extract, Transform, Load (ETL) process to move data from the runtime staging tables to the Unica Campaign contact and response history tables.
  - Logging
    Make sure the log level is set to INFO or ERROR. Never use a verbose log setting like DEBUG or TRACE in a production environment.

Supporting a greater number of concurrent sessions in Ignite

When you are using Ignite as the cache manager in some Unica Interact environments, a high number of concurrent sessions may cause the Unica Interact runtime to exceed its available memory, causing a system slowdown or out-of-memory error. An out-of-memory situation is more likely if you have increased the maxEntriesInCache configuration parameter (Unica Interact > cacheManagement > Caches > InteractCache > maxEntriesInCache) to a number higher than the default setting, although it can occur even with the default setting of a maximum of 100,000 sessions. To avoid this issue, you can reduce the maxEntriesInCache value, or you can follow the instructions here to modify the system memory cache to roll the cached memory over to disk storage. This modification allows many more concurrent sessions than would otherwise be possible.

To prevent the Unica Interact runtime from exceeding the available memory in the Java™ virtual machine (JVM) memory heap, you can modify the memory caching mechanism to use disk storage for caching the data that exceeds the available memory.

However, enabling disk storage will cause a performance degradation. System administrators can adjust the amount of memory available to the deployment systems via the following JVM parameters:

-Xms####m -Xmx####m -XX:MaxPermSize=256m

where the characters #### should be 2048 or higher (depending on their system load.)

Note: A 64-bit application server and JVM are usually necessary for values greater than 2048.

Have feedback?
Google Analytics is used to store comments and ratings. To provide a comment or rating for a topic, click Accept All Cookies or Allow All in Cookie Preferences in the footer of this page.