Distributed cache is a cache sharing solution provided by ByteHouse Cloud Data Warehouse Edition. By leveraging distributed cache technology, it can effectively improve data query performance and meet your high-performance query requirements. This article introduces how to enable and use the distributed cache functionality.

Overview #

ByteHouse Cloud Data Warehouse Edition adopts a storage-compute separation architecture. In real-time data warehouse scenarios, using the local disk cache of the compute group still has certain limitations:

• When relying on object storage and other systems, first-byte latency can affect query response speed;
• The shared storage layer is easily affected by resource competition in multi-tenant scenarios, resulting in performance fluctuations;
• Under the storage-compute separation architecture, if the local cache is not hit (cache miss), it is necessary to directly read from remote storage, which significantly increases query time.

Distributed cache can be deployed across compute nodes in a distributed manner, enabling globally shared cache resources to specifically address the above issues and further enhance performance and stability:

• Data timeliness assurance: Real-time data warehouse business writes are visible within seconds, ensuring data timeliness;
• Tenant-level resource isolation: Provides tenant-level compute and IO acceleration capabilities;
• Adapt to business growth: Cache nodes can be independently expanded, achieving linear IO throughput scalability and flexibly responding to business traffic growth.

After enabling distributed cache, ByteHouse will automatically issue cache configuration for the compute groups in the current environment:

• During queries, data will be automatically loaded into the distributed cache;
• Subsequent identical queries will preferentially read data from the distributed cache, without repeatedly accessing remote storage, significantly improving query efficiency.

Application scenarios #

Distributed cache is applicable to the following scenarios:

• High-performance query scenarios: Suitable for data analysis queries that require rapid response (such as real-time reports and instant data analysis), especially for frequently accessed data. Caching can accelerate queries and reduce query latency.
• Large-scale data processing: In environments with large volumes of data (such as PB-level data storage), distributed caching reduces reliance on remote storage, avoids performance bottlenecks caused by repeated remote data reads, and improves overall query efficiency.

Billing description #

• Pricing and billing:
  • Distributed cache supports unified dedicated cache purchase and management at the "tenant-environment" level. All virtual warehouses share cache resources. Each environment can only enable distributed cache once. After enabling, all users share the cache.
  • The minimum capacity for distributed cache is 10,500 GiB. The annual/monthly package billing mode is supported. For billing details, please refer to Distributed caches.
• Settlement/Renewal, Overdue, Unsubscription: The settlement, renewal cycle, and other logic for distributed cache are independent from the corresponding logic of public virtual warehouses, but the logical framework is similar. For details, see the product billing documentation.

This feature is in Beta. To use it, submit a ticket or contact the ByteHouse team to obtain allowlist privileges.

• To enable the distributed cache feature in the console, use a user with the AccountAdmin or SystemAdmin role to activate the service.
• All sub-users can view and use distributed cache.
• For more information about roles and users, see Role Management.

Enabling distributed cache #

1. Log in to the ByteHouse Cloud Data Warehouse console, click the top Tenant management, click the Basic information tab, in the Distributed Cache module, click Open now to go to the enablement page.
   

2. On the enablement page, set the required cache size, purchase duration, and whether to enable automatic renewal.
   

   Parameter items	Configuration description
   Cache size	The cache size determines the resource capacity of the distributed cache and affects the caching capability for high-frequency access data. The minimum order size is 10500 GiB, and it cannot be reduced. The adjustment increment is 3500 GiB, and the cache specification can be adjusted as needed.
   Purchase duration	Refers to the service validity period, which can be selected as needed.
   Automatic renewal	After activation, the system will automatically pay every month by default, and manual payment is not required. You can adjust the renewal settings at any time. For more information about automatic renewal, see Automatic Renewal.

3. After configuration is complete, click confirm the order. ByteHouse will create the distributed cache and redirect you to the cost center, where you can confirm the cost again and click Next to confirm and pay the relevant fees.

4. After payment is complete, click Go to Console. The system will jump to the ByteHouse Cloud Data Warehouse Console tenant management page, where you can view the initialization status of the distributed cache on the basic information tab. The activated distributed cache will begin initialization, and its status will be "Configuring". The entire initialization process takes about 5 minutes. Please wait patiently.

5. Once the distributed cache initialization is complete, the status changes to "Running".
   

6. (Optional) Edit the project or tag to manage sub-account information. ByteHouse Cloud Data Warehouse Console supports managing sub-account bills based on project or tag. For operation details, please refer to Manage profit sharing.

Expand the distributed cache #

1. In the ByteHouse Cloud Data Warehouse Console, click the top Tenant management, click the Basic information tab, in the Distributed Cache module, click Edit.
   
2. On the edit page, adjust the cache size. The adjustment increment is 3500 GiB. Then click Confirm Change. The system will jump to the order confirmation page, where you can confirm the specifications and fees again. After confirming they are correct, click Next. After payment is successful, the system will automatically expand the cache.
   
3. Click Go to Console. The system will jump to the ByteHouse Cloud Data Warehouse Console tenant management page, where you can view the status of the distributed cache on the basic information tab. During expansion, the distributed cache status displays as Updating. The entire configuration process takes about 5 minutes. Please wait patiently.
   
4. Once expansion is complete, the status changes to "Running", and the cache size is updated to the value you configured.

Manage renewal #

1. In the ByteHouse Cloud Data Warehouse Console, click the top Tenant management, click the Basic information tab, in the Distributed Cache module, click Renew.
   
2. Copy the ID in the prompt box. The following steps require using this ID to search and complete renewal in the BytePlus Cost Center. After copying, click Confirm. The system will jump to the renewal management page of the BytePlus Cost Center.
   
3. Enter the ID copied in Step 2 in the search box after the instance ID, and press Enter to search. Click the Renewal button in the Actions column of the instance row to jump to the renewal information confirmation page. For renewal operations, refer to the Manual renewal rules description. To enable automatic renewal, refer to the Automatic renewal rules description.
4. After renewal is complete, return to the ByteHouse Cloud Data Warehouse Console and refresh the page. The expiration time will be updated according to the renewal period you configured. If you have questions about the renewal period, submit a ticket and contact the after-sales team for processing.

1. In the ByteHouse Cloud Data Warehouse Console, click the top Tenant management, click the Basic information tab, in the Distributed Cache module, click Unsubscribe, and then click Confirm.
   

   warning

   After unsubscription is complete, the existing distributed cache will be emptied. Please proceed with caution.

2. The system will jump to the BytePlus Cost Center's unsubscription management page. Click Submit Order, and then click Confirm to unsubscribe the current resource. After successful unsubscription, the refund amount will be returned to the account balance.
3. After unsubscription is complete, return to the ByteHouse Cloud Data Warehouse Console tenant management page. You can view the state of the distributed cache in the basic information tab. At this time, the state of the distributed cache is "Deleting". The entire deletion process takes approximately 5 minutes.