Amazon Redshift DC2 migration strategy with a buyer case research

This can be a visitor put up by Satoru Ishikawa, Options Architect at Classmethod in partnership with AWS.

In April 2025, AWS introduced the deprecation of Amazon Redshift DC2 cases, guiding customers emigrate to both Redshift RA3 cases or Redshift Serverless. Redshift RA3 cases and Serverless undertake a design that separates storage and compute, affords new options similar to knowledge sharing, concurrency scaling for writes, zero-ETL , and cluster relocation.

On this put up, we share insights from one in all our clients’ migration from DC2 to RA3 cases. The client, a big enterprise within the retail trade, operated a 16-node dc2.8xlarge cluster for enterprise intelligence (BI) and ETL workloads. Going through rising knowledge volumes and disk capability limitations, they efficiently migrated to RA3 cases utilizing a Blue-Inexperienced deployment strategy, attaining improved ETL question efficiency and expanded storage capability whereas sustaining price effectivity.

Amazon Redshift structure sorts

Amazon Redshift affords two deployment choices: Provisioned mode, the place you select the occasion kind and variety of nodes and handle resizing as wanted, and Redshift Serverless, which routinely provisions knowledge warehouse capability and intelligently scales the underlying sources. The next diagram compares these two structure sorts.

Provisioned clusters require you to find out cluster measurement prematurely, however you’ll be able to optimize prices by buying Reserved Cases (RI) or scheduling pause and resume actions. Serverless routinely provisions sources as wanted, with a pay-per-use mannequin the place you solely pay for compute sources consumed. Each companies help migration between one another and provide the identical options together with SQL, zero-ETL, and Federated Question capabilities. For particular pricing particulars, see Amazon Redshift pricing.

Provisioned clusters are appropriate for large-scale, predictable workloads and provide automated scaling based mostly on queuing. Serverless gives management-free automated scaling for variable workloads with AI-driven optimization that scales based mostly on workload complexity and knowledge volumes. For extra particulars, discuss with Evaluating Amazon Redshift Serverless to an Amazon Redshift provisioned knowledge warehouse.

Buyer case research: Migration from DC2 cases

This part describes the client’s migration from Amazon Redshift DC2 to RA3 occasion sorts. The migration used a Blue-Inexperienced deployment strategy that minimized downtime whereas attaining each price optimization and efficiency enchancment.

The client’s workload had the next traits:

Use circumstances

The client had the next key use circumstances for his or her Amazon Redshift deployment:

1. Question by way of BI instrument throughout enterprise hours
   1. Excessive quantity of learn queries
   2. Peak entry throughout Mondays and starting of months
2. Information processing in early morning
   1. Concentrated write queries for knowledge loading and transformation
3. Regular-state workload traits
   1. Run queries greater than 16 hours every day

Necessities

The client had the next key necessities for his or her Amazon Redshift migration:

1. Efficiency
   1. Use auto-scaling (similar to concurrency scaling) throughout peak entry durations
2. Information measurement
   1. Disk capability growth wanted
3. Value Administration
   1. Simple finances prediction and administration
   2. Make the most of low cost companies for long-term utilization
4. Compatibility
   1. Preserve compatibility with current functions and BI instruments
   2. Keep away from endpoint adjustments
5. Availability
   1. Most downtime of 8 hours acceptable throughout migration
6. Community
   1. Don’t modify the prevailing 2-Availability Zone (AZ) subnet configuration
7. When emigrate
   1. To be performed throughout low-load days and hours
   2. Deliberate downtime attainable inside 8 hours

Key concerns in system design, implementation, and operation included prolonged operation hours, ease of finances prediction and administration, price optimization by Reserved Cases (RI), and sustaining compatibility with current methods (avoiding endpoint adjustments). The client evaluated Amazon Redshift Serverless, which supplied engaging options similar to a pay-per-use mannequin, automated scaling capabilities, and the potential for higher value efficiency for variable workloads. Whereas each Redshift Serverless and provisioned clusters might successfully help their workload patterns, the client selected the provisioned mannequin with RA3 nodes, leveraging their years of operational expertise with provisioned environments, current RI technique, and established capability planning strategy.

Options of RA3 occasion kind

Constructed on the AWS Nitro System, RA3 cases with managed storage undertake an structure that separates computing and storage, permitting unbiased scaling and separate billing for every element. These cases use high-performance SSDs for warm knowledge and Amazon S3 for chilly knowledge, offering ease of use, cost-effective storage, and quick question efficiency. For extra particulars, discuss with Amazon Redshift RA3 cases with managed storage.

Migration stipulations

The client had the next migration stipulations in place:

• The client used a Redshift cluster with 16 nodes of dc2.8xlarge configuration.
• The client selected a Blue-Inexperienced deployment strategy for migration, the place they’d restore from a snapshot to RA3 occasion kind, enabling fast rollback if vital.
• The client carried out cluster switching and rollback by endpoint switching utilizing cluster identifier rotation.
• Moreover, to enhance efficiency with excessive concurrency, they transitioned the transaction isolation stage from SERIALIZABLE ISOLATION to SNAPSHOT ISOLATION.

Cluster migration strategies

There have been two migration choices out there: Elastic Resize and Traditional Resize.

Amazon Redshift’s Traditional Resize performance had been enhanced, for resizing to RA3 occasion sorts, considerably lowering the write-unavailable interval. Primarily based on PoC testing, after initiating the resize, the cluster’s standing was modifying for 16 minutes earlier than it turned out there. Primarily based on these outcomes, the client proceeded with the Traditional Resize strategy.

Cluster sizing

Sizing concerned figuring out the occasion kind and variety of nodes for the migration goal. Sizing factors thought of workload traits similar to CPU-intensive (queries utilizing excessive CPU), I/O-intensive (queries with excessive knowledge learn/write), or each.When migrating from DC2 occasion sorts, extra nodes may be required relying on workload necessities. Nodes had been added or eliminated based mostly on the computing necessities for vital question efficiency.

Evaluating configurations with related cluster prices by way of occasion measurement and depend, for a dc2.8xlarge 16-node cluster, the really useful configuration was 8 nodes of ra3.16xlarge. The next was the associated fee comparability within the Tokyo Area:

1. Really helpful: dc2.8xlarge 16-node cluster => ra3.16xlarge * 8-node cluster
   1. $97.52/h (6.095/h * 16 nodes) => $122.776/h (15.347/h * 8 nodes)
2. Value-focused: dc2.8xlarge 16-node cluster => ra3.16xlarge * 6-node cluster
   1. $97.52/h (6.095/h * 16 nodes) => $92.082/h (15.347/h * 6 nodes)

For this migration, the client proceeded with a cost-efficient 6-node ra3.16xlarge cluster to remain inside current finances constraints. Nonetheless, since this node depend might face throughput limitations throughout sure occasions, they enabled concurrent scaling for the RA3 occasion kind to deal with spike entry.

Concurrency scaling gives as much as 1 hour of free credit per day for every energetic cluster, accumulating as much as 30 hours. On-demand utilization charges apply when exceeding this free tier.Whereas the client selected to implement concurrency scaling, Elastic Resize to quickly enhance nodes throughout peak masses was additionally thought of however rejected attributable to on-demand prices for added nodes and the transient disconnection interval throughout switching.

Managed storage price

RA3 cases use Redshift Managed Storage (RMS), which is charged at a hard and fast GB-month fee. The client’s roughly 2 TB of information required together with storage prices within the estimates. For pricing particulars, see Amazon Redshift pricing.

Migration step from DC2 to RA3

After creating an RA3 cluster from the DC2 cluster’s snapshot, the client swapped the cluster identifiers. The next diagram reveals this course of.

1. Take a snapshot of the present DC2 cluster.
2. Restore RA3 cluster from the snapshot with a special cluster identifier (Traditional Resize)
3. Swap the cluster identifiers between the present DC2 cluster and the brand new RA3 cluster.

If any points come up after the cluster swap, you’ll be able to shortly roll again by returning the unique DC2 cluster to its authentic cluster identifier.

Word: Restore from a snapshot

Operating the restore operation utilizing CLI instructions is really useful to attenuate operational errors and guarantee reproducibility. The next is a pattern command.

aws redshift restore-from-cluster-snapshot 
--cluster-identifier for-ra3-20250207 
--snapshot-identifier cm-cluster-for-ra3-20250207 
--cluster-subnet-group-name cm-cluster 
--vpc-security-group-ids sg-1234567a sg-2345678b sg-3456789c 
--cluster-parameter-group-name cm-cluster 
--node-type ra3.16xlarge 
--number-of-nodes 6 
--port 5439 
--no-publicly-accessible 
--enhanced-vpc-routing 
--availability-zone ap-northeast-1a 
--preferred-maintenance-window sat:17:00-sat:17:30 
--automated-snapshot-retention-period 14 
--iam-roles 'arn:aws:iam::123456789012:function/AmazonRedshift-CommandsAccessRole' 'arn:aws:iam::123456789012:function/AmazonRedshift-Spectrum' 
--maintenance-track-name present

Manufacturing migration period

The time required for the restore and traditional resize steps can differ considerably relying on knowledge quantity and goal cluster specs. The client performed a rehearsal beforehand to measure the precise required time.

Check outcomes

Earlier than the manufacturing migration, the client created a take a look at cluster by restoring a snapshot to the RA3 occasion kind. Whereas Redshift Check Drive is usually helpful for workload testing, this buyer confronted distinctive constraints: enabling audit logging of their manufacturing cluster would require configuration adjustments, cluster restarts, and complicated approval processes beneath their strict change administration insurance policies. To deal with this, they developed a customized load testing instrument that captured workload patterns utilizing Amazon Redshift system views (SYS_QUERY_HISTORY and SYS_QUERY_TEXT), which keep 7 days of question historical past. The instrument replayed 55,755 historic queries with 50-way parallelism in opposition to each DC2 and RA3 clusters, evaluating metrics together with question execution time, CPU utilization, and disk I/O. Question outcome caching was disabled throughout testing to make sure correct comparisons.

BI question efficiency

BI queries had been examined utilizing the customized load testing instrument. The outcomes symbolize the typical execution time from 15 take a look at runs of 55,755 queries executed with 50-way parallelism. With out concurrency scaling, the dc2.8xlarge 16-node cluster averaged 45.82 seconds per question, whereas the ra3.16xlarge 6-node cluster averaged 91.30 seconds. This indicated that RA3 cases confirmed longer execution occasions for brief and medium queries in a direct migration with out optimizations. Nonetheless, enabling concurrency scaling improved RA3 efficiency progressively. With concurrency scaling enabled at most 2 clusters, the ra3.16xlarge 6-node cluster achieved a median of 72.48 seconds per question, a 21% enchancment over the non-scaled configuration.

ETL question efficiency comparability

For long-running ETL queries (execution time higher than 10 minutes), the RA3 cluster demonstrated higher efficiency than DC2. These outcomes represented a direct migration of the client’s workload with no optimizations utilized.

• For the Massive-scale knowledge load workload 1, the ra3.16xlarge cluster accomplished the question 28% quicker than the dc2.8xlarge cluster (41 minutes vs. 57 minutes).
• For the Advanced transformation workload 1, the ra3.16xlarge cluster was 23% quicker (1 hour 1 minute vs. 1 hour 20 minutes).

These outcomes indicated that the RA3 node kind was extra performant for time-intensive knowledge loading and transformation duties. The upper CPU utilization values for RA3 recommended simpler compute useful resource utilization.

Efficiency tuning

Primarily based on the take a look at outcomes, the client recognized that RA3 confirmed longer execution occasions for brief and medium BI queries however quicker efficiency for long-running ETL queries in comparison with DC2. To optimize total efficiency, they centered on figuring out sluggish queries and continuously referenced tables, prioritizing optimizations with the very best impression.

Efficiency tuning technique

The client thought of a number of optimization methods to leverage RA3’s architectural benefits. One key technique concerned pre-processing ad-hoc brief and medium question workloads throughout low-load durations, creating pre-processed tables or materialized views for queries that repeatedly carried out joins, aggregations, filters, and projections. RA3’s separated compute and storage structure, with cost-effective large-scale storage, supported this strategy.

Changing common views to materialized views

Evaluation of sluggish queries revealed the usage of joins in views, and continuously referenced tables had been being accessed a number of occasions by these views. As a countermeasure, the client changed continuously used common views with materialized views, eradicating pointless knowledge ranges and redundant columns.

Amazon Redshift helps incremental updates of materialized view contents by way of the REFRESH MATERIALIZED VIEW command, enabling environment friendly knowledge updates.

Materialized views and question rewrite

By changing common views to materialized views, current queries could also be routinely optimized by the “question rewrite” function supplied by the question planner. For extra particulars, discuss with “Computerized question rewriting to make use of materialized views“.

Computerized tuning with AutoMV

On the DC2 cluster, disk utilization constantly exceeded 80%, which disabled the AutoMV function attributable to inadequate disk house. With RA3’s expanded storage, automated tuning by AutoMV turned attainable, resulting in additional efficiency enhancements. For extra particulars about AutoMV, discuss with Automated materialized views.

Efficiency tuning outcomes

After making use of these optimizations, the client achieved the next outcomes:

• Maintained current efficiency whereas controlling price will increase
• Achieved larger CPU utilization whereas sustaining throughput
• Enhanced dynamic throughput throughout peak load durations utilizing concurrency scaling’s automated scaling

Conclusion

On this put up, you discovered how a big retail enterprise efficiently migrated from Amazon Redshift DC2 to RA3 cases. The Blue-Inexperienced deployment strategy enabled a protected migration with fast rollback functionality, whereas the separated compute and storage structure of RA3 supplied flexibility to deal with rising knowledge volumes. Though RA3 confirmed completely different efficiency traits for brief BI queries in comparison with DC2, the client achieved vital enhancements in long-running ETL question efficiency (as much as 28% quicker for knowledge masses and 23% quicker for complicated transformations). By leveraging RA3-specific options similar to materialized views and AutoMV, they optimized total question efficiency whereas sustaining price effectivity by Reserved Cases and concurrency scaling.

To proceed your RA3 migration journey, see Greatest practices for upgrading from Amazon Redshift DC2 to RA3 and Amazon Redshift Serverless and Resize Amazon Redshift from DC2 to RA3 with minimal or no downtime for added steering and greatest practices.

In regards to the authors