A theme runs through everything this month: how fast enterprise data can actually move when nothing is standing in its way.
- 12–14 hours down to 2–4 — a Fortune 500 retailer’s final data migration, with 95% staged in advance
- 60.9 million records in 18 minutes — Marc and Vince on why ETL and change data capture run out of room at scale
- Months of mainframe backups for pennies — how we protect our own mainframe with FlowZ
- Core banking data, reachable for AI — a new joint white paper with Zafin
None of these are projections. They’re what happened in production this summer, in customer environments and in our own.
Let’s get into it. ↓
Customer White Paper
New Joint White Paper with Zafin: Banking AI Needs Core Data
Some of the data most valuable to banking AI — transactions, balances, customer history, operational context — still lives on the mainframe in formats modern tools can’t read directly.
Our new joint white paper with Zafin shows how Zafin IO and VirtualZ PropelZ work together to decode that data, move usable copies into modern environments, and make it directly queryable — while IBM Z stays the trusted system of record.
Inside: how PropelZ reduces the dependence on hand-built COBOL copybooks, how Zafin opens decoded data up for reconciliation, analytics, and AI-assisted investigation, and how one top-tier bank’s investigation workflow went from an hours-long, multi-tool process to opening and querying multi-million-row files in minutes.
No core migration. No application rewrite.
Customer Case Study: The Metrics
When the Transfer Window Is 10 Hours but the Transfer Takes 14
A Fortune 500 US retailer is moving 125 applications onto Google Cloud. Cutover is a single, unforgiving event — VSAM, GDGs, flat files, historical archives, and live transactional change all have to be current to the second when processing shifts.
The client’s standard Managed File Transfer path put the final pre-cutover migration at 12 to 14 hours, against a downtime target of 8 to 10. The transfer alone would have consumed the window.
After a six-month proof of concept, the customer selected PropelZ. Compression collapsed the transfer window, and change data capture replicated ongoing change so the bulk of the work happened ahead of cutover instead of inside it. The final migration now runs in 2 to 4 hours, with roughly 95% of the data staged in advance. Along the way: 60.9 million records in 18 minutes at a sustained 56,000 records per second, and 460,000 VSAM records into PostgreSQL on GCP in 25 seconds — single step.
Validated across VSAM, non-VSAM, GDG, and tape datasets in the 100–200M+ record range, with data integrity confirmed on every transfer and no impact to production workloads.
New on the Blog
What’s the Best Way to Do Mainframe Backup and Archive to the Cloud?
We run a mainframe, which means we have the same quiet problem every mainframe shop has: data that has to stick around, stay cheap, and survive losing the building.
So we pointed our own backups at cloud object storage with FlowZ and stopped thinking about it. This post is the honest accounting — weekly full-volume system backups, daily source-code backups, and SMF log archival running in production for over a year. What it costs (pennies for months of retention). How fast it actually is (a 3390-54 in about two minutes over a plain gigabit line). And why setup was a one-time config exercise rather than a project.
No custom code. No cloud-side software. And because FlowZ is zIIP-eligible, minimal impact to your general-purpose MIPS.
Press Play
Now Streaming on the VirtualZ YouTube Channel
One channel. Every angle. Start here.
Our curated playlists walk you through AI & Modernization Strategy, Use Cases in Action, Product Deep Dives, Security & Compliance, Strategic Partner Conversations, and more — in the order we recommend watching.
Secure, Connected Enterprise Data — Built for What’s Next
If you’ve ever needed a short, clear way to explain what VirtualZ does — to a colleague, a stakeholder, or someone new to your modernization team — this is it.
AI has become a competitive differentiator, and success depends on data people can actually trust and reach. This video shows how VirtualZ delivers secure, connected enterprise data across IBM Z, cloud platforms, and hybrid environments — supporting AI, analytics, modernization, secure file exchange, and virtual cloud storage.
On the Skyward Data Podcast
Why ETL and Change Data Capture Break at Mainframe Scale
Sixty million VSAM records. From z/OS to PostgreSQL in the cloud. A sustained 56,000 records per second — no staging database, no change data capture, no custom code. The same file on the retailer’s previous solution ran for two days.
On this episode of Skyward Data, Executive Chairman Marc Sokol and CTO Vince Re walk through how that actually happened, and why the three traditional paths — hand-built ETL scripts, CDC deployments that require VSAM logging and RLS, and staged gateway architectures that read and write the same data four times — run out of room at enterprise scale.
Also in the episode: the real benchmark environment (a mid-sized z/OS system, 4K+ records, a shared 10 Gb link, GCP-hosted PostgreSQL — using 3–5% of available bandwidth), how COBOL copybook metadata drives field-level transformation declaratively, why JDBC bulk operations beat per-record web services, and what VSAM logging actually costs you in I/O, journal storage, and batch window.
Skyward Data is THE podcast for professionals who want to gain a competitive edge in the evolving data landscape. Join us on your favorite platform, and subscribe to remain ahead of the competition.
Any Platform. Any Input. Any Output. One Patented Portfolio.
On-Demand Customer Briefings
Today’s enterprise leaders want insight, not a sales pitch. That’s why we created the VirtualZ Customer Briefing Center — a self-paced, no-pressure environment where you can explore our patented, no-code data access software when and how it works for you.
“These briefings are designed to put customers in control. They make it easy to explore how our solutions solve real-world problems — at your own pace, with clarity and no noise.”
— Vince Re, CTO, VirtualZ Computing
Thank you!
Thanks for spending a few minutes with us this month.
Whether you’re modernizing storage, accelerating AI, or powering compliance, VirtualZ is here to help you unlock mainframe data with less friction — with our no-code solutions, you can start delivering results today.
Ready when you are:
- Explore at your own pace in the Customer Briefing Center
- Watch demos and webinars anytime on YouTube
- Reach out directly at hello@virtualzcomputing.com
See you next month with more ways to modernize, integrate, and accelerate.
— The VirtualZ Team