Performance Options for the Connect Capability of TIBCO Cloud Integration

Integration performance presents a unique situation for every project and customer environment. Even the most detailed benchmarks in your integration plan can be challenging in many scenarios, due to the use case being different than your project. There is no magic formula on how long an integration or migration will take, however, TIBCO Cloud Integration presents you with many tools at your disposal to address performance and expedite the process.

Elements that impact performance:

A few of the most common situations that can negatively impact integration performance include:

• A large number of lookups and data transformations
• Network latency
• Source and target application performance

A solid understanding of elements that can influence performance leads you to the question ?What changes can be made to positively impact and improve performance?? TIBCO Cloud Integration empowers you with the ability to improve performance in a multitude of ways. The most common and simple performance enhancement one can make is:

Taking advantage of the agent architecture, which allows you to distribute your integration maps across various solutions instead of placing all your integration maps into a single solution.

Prior to getting started, here are a few definitions you will want to know:

Solution

• A Solution is a set of user-specified configuration options that TIBCO Cloud Integration uses to perform a specific task, such as replicating or integrating entities or fields from a source datastore to a target datastore.
• For more information click here 

Maps

• A Map is a visual representation of instructions needed to integrate your data. Sets of one or more TIBCO Cloud Integration Maps are used in Integration and Migration Solutions to retrieve and manipulate data.
• For more information click here 

Agents

• The TIBCO Cloud Integration Agent facilitates communication between the data sources used by a Solution. It provides secure communications to TIBCO Cloud Integration without opening security holes in your corporate firewall. An Agent is required to communicate with both your source and target data.

  There are two types of Agents available:

  • An On-Premise Agent-This Agent is installed on a computer at your site. You can install multiple On-Premise Agents on a single computer.
  • A Cloud Agent-This Agent resides in the cloud. 
  • For more information click here 

Performance Architecture Using an On-Premise Agent:

• Agents are multi-threaded and can take advantage of their host systems resources i.e. CPUs and memory. Keeping this in mind, it is still a recommended approach to take advantage of the agent's flexibility with large integration projects.

• Keeping your business process in mind, architect your integrations in a way that allow you the flexibility to distribute different workloads needed for the project across multiple solutions.

Example:

• Solution 1 ? Group high transactional integration maps which require a more frequent schedule.
• Keep high priority data separate for greater reporting frequency.

• Solution 2 ? Group less transactional integration maps which don?t require a frequent schedule.
• Keep lower priority data integration away from higher priority data.

• Solution 3 ? Group in bulk, any large data volume integration maps containing entities which have no modify date and require a much longer run time for every scheduled run.
• Keep lesser needed reporting and business data away from higher priority jobs to help mitigate delays.

Increase processing power and shorten integration time by deploying a different agent, across different on-premises systems.

Having a different agent per solution allow for the opportunity to process its own resources without impacting others.

Figure 1 example of on-premises agent scaling

Performance Architecture Using a Cloud Agent:

• When using a cloud agent as part of your performance architecture, the same solution grouping is applicable, the only caveat being you can only manually install one.

• Despite only being able to only deploy one cloud agent manually, you will only see the one Cloud Agent in your Org. It will appear as though you are assigning the same agent to various solutions. However, the Scribe cloud will be auto scaling agents to support a Solution that you assigned it to.
• In effect the feature is creating redundancy and additional multi-processing for your additional integration solutions.

Figure 2 example of cloud agent scaling

Summary:

We covered one of the most direct and simple ways to address your integration performance needs using out of the box tools in this blog which included:

• Using multiple solutions in your project you can then assign individual agents to each taking advantage of additional processing power to maximize your integration performance.
• Architecting and distributing different integration workloads based on your business process by using multiple solutions.

First published April 04, 2017