# Replica v3 Data Delivery Platform Upgrade

## Summary

Peach is upgrading its data delivery infrastructure to Replica v3, a more reliable, scalable, and flexible platform. This upgrade transitions clients from direct database access to file-based delivery with modern data formats and improved delivery consistency.

## Timeline

- **Availability:** Replica v3 is available now
- **Migrations Begin:** Q1 2026
- **Parallel Operation:** Both legacy and v3 systems run simultaneously during your transition
- **Legacy Deprecation:** March 30, 2027


## What Is Changing

Your current database replica access is being replaced with a modern file-based delivery system.

| Aspect | Current (v1/v2) | New (v3) |
|  --- | --- | --- |
| Data Delivery | PostgreSQL database connection | File-based delivery (Parquet) |
| Delivery Time | Variable; no guaranteed completion | Options: Daily (once at 6 AM local), Twice Daily (6 AM and 6 PM local), Hourly |
| Frequency Options | Daily only | Daily, Twice Daily, or Hourly |
| Scalability | Performance degrades above ~25K active loans | Scales efficiently to 1M+ portfolios |
| Data Format | PostgreSQL tables | Parquet files |
| Access Method | Direct database queries | Cloud storage bucket (push or pull) |


## Why We Are Making This Change

- **Reliability:** The current system is susceptible to failures that delay data availability. The new architecture provides more consistent delivery.
- **Scalability:** Replica v1/v2 cannot efficiently support portfolios above 75,000 loans. v3 scales to over 1 million loans without performance issues (less than 20 minutes to replicate).
- **Reduced Processing:** Modern formats (Parquet) require less transformation and integrate directly with data warehouses like Snowflake, Databricks, and BigQuery.
- **Flexibility:** New frequency options (twice daily, hourly) enable faster access to data for credit, risk, and operational reporting.


## Updated Pricing

| Delivery Option | Price (per 25K loans) | Delivery Schedule |
|  --- | --- | --- |
| Daily File | $140/month* | Once daily |
| Twice Daily File | $160/month | Twice daily |
| Hourly File | $210/month | Hourly updates (up to 24/day) |


*First 25,000 active loans included at no additional cost for Daily File.

## Migration Guide

### Prerequisites

- Cloud storage bucket (AWS S3, Google Cloud Storage, or Azure Blob Storage)
- Service account credentials with appropriate permissions
- Resources available for ETL pipeline updates
- Completed contract amendment with Customer Success


### Migration Steps

**Step 1: Coordinate with Peach**

- Customer Success will contact you to initiate the migration process
- Complete any required contract updates
- Select your preferred file format (Parquet) and delivery frequency


**Step 2: Configure Cloud Storage Access**

Choose one of two delivery options:

- **Option A (Push):**
  - Peach pushes files to your cloud storage bucket
  - You provide Peach with service account credentials for write access
  - Peach configures automated delivery to your specified location
- **Option B (Pull):**
  - You pull files from a Peach-hosted storage bucket
  - Peach provides read access credentials to a Google Cloud Storage bucket
  - You configure your ETL pipelines to pull files on your preferred schedule
- **Option C (Snowflake Reader Accounts):**
  - Access your data via a Peach-provided Snowflake Reader Account using Snowflake Secure Data Sharing
  - See [Access Peach Data w/ Snowflake Reader Accounts](/data-reporting/snowflake-reader-accounts) for setup instructions
- **Option D (Direct Share):**
  - Peach shares your data directly to your existing Snowflake account via Snowflake Secure Data Sharing
  - See [Access Peach Data w/ Snowflake Direct Share](/data-reporting/snowflake-direct-share) for setup instructions


> **Note:** Detailed setup steps for specific cloud platforms (AWS, Azure, GCP, Databricks, Snowflake) are available in the [Technical Migration Guide](/release-notes/change-notices/replica-v3-migration-guide).


**Step 3: Update Your Data Pipelines**

- Modify ETL processes to read from cloud storage instead of the PostgreSQL database
- Add parsing logic for Parquet file format
- Update scheduling as needed based on your selected delivery frequency


**Step 4: Validate and Test**

- Both v1/v2 and v3 systems run in parallel during your migration
- Compare data from both sources to validate completeness and accuracy
- Test all downstream processes with v3 data before cutover
- Notify Peach when validation is complete


**Step 5: Cutover**

- Coordinate final cutover date with your Peach contact
- Switch production workloads to v3 data source
- Confirm successful transition to begin decommissioning legacy access


## Frequently Asked Questions

**Will there be any downtime during the migration?**

No. Both legacy and v3 systems operate simultaneously during your migration. You can continue using your current data access while testing the new system. There is no expected service interruption.

**Which file format is supported?**

Parquet. This format is natively supported by Snowflake, Databricks, BigQuery, AWS Athena, and most modern ETL tools.

**What if I have a sandbox environment?**

If you have replica access in sandbox, it will be migrated concurrently with production. Both can be done in parallel.

**What happens to my Campaigns functionality?**

Campaigns have been tested and validated with the new infrastructure. Your existing campaign queries will continue to work. Peach will assist with any campaign-related migration tasks.

**When will the legacy replica be decommissioned?**

March 30, 2027

## Next Steps

Customer Success will contact you to:

1. Discuss your specific migration timeline and requirements
2. Review pricing and contract updates applicable to your account
3. Coordinate technical migration steps with your team
4. Review the [platform-specific configuration documentation](/release-notes/change-notices/replica-v3-migration-guide)


## Questions

Contact [support@peachfinance.com](mailto:support@peachfinance.com)