SFTP straight into Azure
SFTP endpoints on Azure Data Lake Storage with no servers to patch — a separate login and home folder per sender, SSH keys, and access limited to known networks.
Solutions · Data & Analytics
Managed Data Transfer & Conversion
Files still move through overnight scripts nobody wants to touch, mainframe data arrives in EBCDIC that modern tools can't read, and when a feed fails nobody knows until a report is wrong. Symetriz transfers, converts, verifies, and lands your data in Azure — as a one-off migration or a managed daily feed.
The problem
Still moving files by hand at 2 a.m.?
Fragile overnight scripts
FTP jobs and shell scripts written years ago move business-critical files, and nobody wants to touch them.
How we fix it: Managed, monitored transfers with retries and alerts.
Data modern tools can't read
Mainframe files arrive in EBCDIC, with packed decimals and variable-length records that break standard loaders.
How we fix it: Copybook-driven conversion to clean UTF-8.
Failures nobody notices
A late, missing, or truncated file is only discovered when a report comes out wrong days later.
How we fix it: Reconciliation and alerts on every file, every run.
The solution
Transferred, converted, verified, landed
- Secure SFTP or pipeline transfer into your own Azure subscription
- Copybook-driven EBCDIC-to-Unicode conversion, packed decimals included
- Every file reconciled — record counts and control totals, not hope
Transfer, securely
The first mile is where most feeds break — and where many data leaks start.
Pipeline & partner pulls
Scheduled pulls from FTP, SFTP, file shares, and databases with Azure Data Factory, for sources that can't push — including mainframe datasets and their copybooks.
Encrypted, private, audited
Encryption in transit and at rest, private endpoints for internal traffic, and a record of who sent which file, and when.
Convert, without losing a cent
Mainframe data isn't just text in a different alphabet. Read one field at the wrong offset or type, and numbers corrupt silently.
CUSTMAST.cpy → CUSTMAST.dat → customers (UTF-8)
| Copybook field | EBCDIC bytes (hex) | Lands as |
|---|---|---|
| 05 CUST-NAME PIC X(4). | C1 E2 C8 C1 | ASHAtext |
| 05 OPEN-DATE PIC 9(8). | F2 F0 F2 F6 F0 F9 F2 F9 | 2026-09-29zoned → date |
| 05 BALANCE PIC S9(7)V99 COMP-3. | 12 34 56 78 9C | 1234567.89packed, + |
| 05 ADJUSTMENT PIC S9(3)V99 COMP-3. | 01 25 0D | -12.50packed, − |
Copybook-driven layouts
Your COBOL copybooks drive the conversion, so every field is read at the right offset, length, and type — including REDEFINES and OCCURS DEPENDING ON.
Code pages & character sets
EBCDIC code pages such as 037 and 500 converted to UTF-8, with national characters and special symbols mapped rather than dropped.
Packed & zoned decimals
COMP-3 packed and zoned decimal fields unpacked with their signs and implied decimal points intact, so balances and quantities stay exact.
Verify, then land
Data you can't prove arrived intact is data nobody should build on.
Reconciliation on every file
Record counts, control totals, and checksums compared between source and target, with mismatches stopped before they load.
Land it in the lake
Raw and converted files written to a landing zone in Azure Data Lake Storage, organized by feed and date, ready for your lakehouse or analytics team.
Or load it straight into a database
Converted records loaded into Azure SQL, PostgreSQL, or Microsoft Fabric, with schema mapping, keys, and data types agreed up front.
One-off migration or managed feed
Move a historical archive once, or run the same flow every night — the controls are the same.
One-off bulk migration
Historical archives and tape extracts moved in rehearsed batches, reconciled, and signed off before the source system is retired.
Recurring managed feeds
Daily, hourly, or event-driven feeds watched for late, missing, or malformed files, with alerts and reruns handled for you.
Retention & audit evidence
Retention rules, immutable copies where regulation requires them, and a run history you can hand to an auditor.
Ready when you are
Tell us what's in the files.
Share a sample file, its copybook, and where the data needs to go, and we will scope the transfer, conversion, and landing — starting with a converted sample you can check.