Deployment Scenario to Achieve Scalability for Site Visitor Read Access | Community
Skip to main content
S
S__k__Agarwal
Level 3
April 26, 2025
Solved

Deployment Scenario to Achieve Scalability for Site Visitor Read Access

  • April 26, 2025
  • 2 replies
  • 428 views

To achieve scalability for site visitor read access, which deployment scenario should be used?

 

• Oak Cluster with MongoMK across Multiple DataCenter

• Single TarMK Instance

• TarMK Farm

• Oak Cluster with MongoMK in Single DataCenter

 

@aanchal-sikka @EstebanBustamante 

@arunpatidar @Shashi_Mulugu 

@lukasz-m @Mahedi_Sabuj 

@kautuk_sahni @Sudheer_Sundalam

@lukasz-m @Rohan_Garg 

Best answer by ShivamKumar

Hi @s__k__agarwal ,

I think TarMK Farm is the right answer because It is designed to provide high availability and scalability specifically for read-heavy access in AEM publish environments. In this deployment model, multiple independent TarMK publish instances are deployed behind a load balancer, allowing the system to distribute read traffic efficiently.

 

This setup results in a more responsive and scalable experience for site visitors, especially under high traffic loads.

2 replies

S
ShivamKumarAccepted solution
Level 4
April 26, 2025

Hi @s__k__agarwal ,

I think TarMK Farm is the right answer because It is designed to provide high availability and scalability specifically for read-heavy access in AEM publish environments. In this deployment model, multiple independent TarMK publish instances are deployed behind a load balancer, allowing the system to distribute read traffic efficiently.

 

This setup results in a more responsive and scalable experience for site visitors, especially under high traffic loads.

    partyush
    Community Advisor
    partyushCommunity Advisor
    Community Advisor
    April 26, 2025

    Hi @s__k__agarwal 

     

    Use a TarMK Farm for scalability in site visitor read access. It’s simple, cost-effective, and built for AEM’s publish tier. Set up multiple publish instances, replicate content from Author, and use Dispatcher for caching and load balancing. Scale by adding more publish nodes as needed.

    Implementation Steps:

    1. Set Up Multiple Publish Instances:
      • Deploy multiple AEM publish instances (e.g., publish1, publish2, etc.) on separate servers or containers.
      • Each instance uses its own TarMK repository (stored locally in crx-quickstart/repository).
    2. Configure Replication:
      • From the AEM Author instance, set up replication agents to push content to all publish instances (/etc/replication/agents.author).
      • Example: Create a replication agent for each publish instance in /etc/replication/agents.author/publish1, publish2, etc.
    3. Use Dispatcher for Caching:
      • Place a Dispatcher (Apache with mod_dispatcher) in front of the publish farm.
      • Configure Dispatcher to cache content and load-balance requests across publish instances.
      • Example dispatcher.any config:
        /renders { /publish1 { /hostname "publish1" /port "4503" } /publish2 { /hostname "publish2" /port "4503" } }
    4. Scale Horizontally:
      • As traffic grows, add more publish instances and update the Dispatcher config to include them.
      • Ensure replication agents are updated to push content to new instances.
    5. Monitor and Optimize:
      • Use AEM’s monitoring tools (/system/console/jmx) to track performance.
      • Tune Dispatcher cache settings and TarMK segment store for optimal read performance.

    Here is the link for more info : https://experienceleague.adobe.com/en/docs/experience-manager-65/content/implementing/deploying/deploying/recommended-deploys

    Thanks.

     

    Best 

    Partyush

      Terms & ConditionsAccessibility statement

      Loading footer...