A managed file transfer (MFT) platform is an internet-facing server that holds your most sensitive bulk data and the credentials of every trading partner, which is why file transfer products have been at the centre of several of the largest data theft campaigns of recent years. CISA's Known Exploited Vulnerabilities (KEV) catalogue lists at least 12 exploited flaws in file transfer and FTP server products added since February 2023, six of them flagged as used in ransomware campaigns. Buying MFT is therefore a security decision first and an integration decision second. This vendor-neutral buyer's guide is for security architects, integration teams and procurement leads replacing ad hoc FTP, scripts or a legacy MFT server. It covers requirements, capability areas, evaluation criteria, vendor questions, red flags and a proof of concept plan. For operating an MFT platform once bought, see the Managed File Transfer (MFT) Security Guide; for user-driven sharing through SharePoint, Google Drive or Box, see the Secure File Sharing Guide.
Key takeaways
- Treat the platform as an edge device. Patch speed, a small exposed surface and the ability to put a gateway or reverse proxy in front of the core matter more than the number of connectors.
- Every trading partner is a machine identity. Partner accounts, SSH keys, certificates, PGP keys and API tokens need an owner, an expiry and a rotation process, and the platform should make that visible.
- Do not keep what you do not need. Automatic deletion of files after delivery limits what a single exploit can steal; data extortion campaigns against MFT servers took whatever was sitting on disk.
- Logging must leave the box. File-level audit events, administrative changes and authentication events should stream to your SIEM in a documented format so you can detect bulk downloads and investigate after a vendor advisory.
- Test the vendor, not just the product. Ask how fast past critical vulnerabilities were fixed and disclosed, and whether a hosted service is patched for you before public disclosure.
Define requirements
Inventory existing transfer flows before you look at products. Most organisations find a mix of scheduled scripts, legacy FTP servers, email attachments and partner portals. For each flow record the partner, protocol, direction, data classification, volume, schedule and business owner. That inventory drives the protocol list, the deployment model and the sizing.
| Protocol or interface | Typical use | Security notes for requirements |
|---|---|---|
| SFTP (SSH File Transfer Protocol, TCP 22) | System-to-system batch transfers, most partner integrations | Require modern SSH key exchange and ciphers, key-based authentication, per-partner keys and host key pinning by partners. See the SSH Key Management Guide. |
| FTPS (FTP over TLS, RFC 4217) | Legacy partners that cannot use SSH | RFC 4217 protects the control channel with AUTH TLS and the data channel only after PBSZ 0 and PROT P, because data connections start in clear. Require the server to refuse unprotected data connections, and require TLS 1.2 or later; plan for firewall handling of passive data ports. |
| HTTPS web portal and REST API | Ad hoc human transfers, application integration | Single sign-on for staff, MFA for external users, API tokens scoped per integration with expiry. |
| AS2 (RFC 4130) | EDI in retail, logistics and healthcare | S/MIME signing and encryption with signed message disposition notifications (MDNs) giving proof of receipt; certificate lifecycle needs managing. |
| Cloud object storage connectors | Moving files to and from cloud buckets | Use workload identity or short-lived roles rather than long-lived access keys stored in the MFT configuration. |
| Plain FTP | None for new requirements | Credentials and data cross the network in clear text; plan its retirement as part of the project. |
- Deployment model: self-hosted software gives full control of data location and network placement but makes you responsible for patching within hours of an emergency advisory. A vendor-hosted (SaaS) service moves patching to the provider but places partner credentials and data in its environment, so assurance, data location and breach notification terms become central. Hybrid designs keep the core on premises with a hardened gateway in a DMZ.
- Small vs large organisations: a small organisation with a handful of partners is often safer with a hosted service and strict retention; a large enterprise with hundreds of flows usually needs workflow automation, high availability, segregation of business units and an internal gateway tier.
- Regulatory drivers: record which flows carry personal data, payment card data or health data, because those flows set encryption, retention and audit evidence requirements. See the Data Classification Guide.
Migrating from legacy transfers
Most MFT projects are migrations, and the migration is where weak practice is carried forward. Plan it as a security improvement, not a lift and shift.
- Re-issue, do not copy, credentials: give every partner a new account and a new SSH key or certificate on the new platform rather than importing old passwords. Old credentials may already be in a breach dump or a former employee's notes.
- Move partners up the protocol ladder: use the cutover to move plain FTP partners to SFTP or FTPS with protected data channels, and set a date after which the legacy protocol is switched off.
- Retire scripts that embed secrets: scheduled jobs with passwords in batch files or cron entries should become managed workflows that fetch secrets from a vault at run time.
- Agree the cutover per flow: run old and new routes in parallel for a short, fixed period, compare checksums and delivery receipts, then disable the old route and remove its firewall rules. Keep a documented rollback for each flow until its business owner signs off.
- Decommission properly: wipe the old server's file stores, revoke its certificates and keys, and remove its DNS records so partners cannot fall back to it.
Capability areas
| Capability area | What good looks like | Illustrative weight |
|---|---|---|
| Exposure reduction and architecture | Separate gateway or proxy tier so the core and its database are not internet-facing; minimal listening services; ability to disable unused protocols and web interfaces | 15% |
| Vulnerability response | Published security advisory process, CVE assignment, historical time from report to fix, emergency hotfix path, notification to customers | 15% |
| Partner and credential management | Per-partner accounts, SSH key and certificate inventory with expiry alerts, IP allow lists per partner, MFA for human users, API token scoping | 15% |
| Encryption and key handling | TLS 1.2 or later and modern SSH algorithms by default, encryption of stored files, PGP for end-to-end payload protection, keys in an HSM or external key manager option, FIPS 140-3 validated modules where you need them | 10% |
| Data minimisation | Per-flow retention, automatic deletion after delivery, purge of temporary and staging files, no copies kept by default in web portals | 10% |
| Audit and monitoring | File-level events (upload, download, delete, failed authentication, admin change) exported by syslog or API in a documented schema, integrity of the audit trail, alerting on bulk downloads | 15% |
| Workflow and integration | Scheduling, retries, checksum verification, content scanning hooks (antivirus, DLP), secrets pulled from a vault rather than stored in workflow definitions | 10% |
| Resilience and operations | High availability, tested backup and restore of configuration and keys, documented upgrade path, supported operating systems | 10% |
Where identity fits
An MFT platform is a concentration point for non-human identities. Partner SFTP keys, AS2 certificates, service accounts used by scripts, cloud storage credentials and API tokens often outlive the projects that created them, and a stolen administrator credential or database dump can expose all of them. Ask for an export of every credential the platform holds, with owner, creation date and last use, and plan to rotate partner credentials after any vendor security advisory. See the Machine Identity, PKI and Certificate Lifecycle Guide and the Secrets Management Guide.
Evaluation criteria
| Criterion | Evidence to request | Example pass threshold |
|---|---|---|
| Patch history | List of critical and high CVEs in the last three years, dates reported, fixed and disclosed | Fixes for critical flaws available within days, with clear customer communication and indicators of compromise |
| Secure defaults | Default configuration export from a fresh install | Plain FTP, weak TLS and SSH algorithms, and anonymous access disabled by default |
| Authentication | Configuration screens and documentation | SSO via SAML or OpenID Connect for staff, MFA for portal users, key or certificate authentication for systems, admin interface reachable only from management networks |
| Logging | Sample log events and schema documentation | Every file event carries user or partner, source IP, file name, size, hash and result; logs exported in near real time |
| Data retention | Retention settings per flow | Deletion after successful delivery configurable per flow, including portal and temporary storage |
| Assurance | ISO/IEC 27001 certificate and scope, SOC 2 Type II report for hosted service, penetration test summary, secure development statement | Scope covers the product and hosting operations |
| Exit | Configuration and partner export format | Partner definitions, keys (where policy allows) and workflows exportable for migration |
For a structured scoring sheet, see the Buyer's Guide Methodology and Scoring Template.
Questions to ask vendors
- What was your last critical vulnerability, how was it reported, how many days to a fix, and how did you notify customers who run the software themselves?
- Which components are internet-facing in your recommended architecture, and can the database and admin interface be kept off the internet entirely?
- How do you store partner passwords, private keys and certificates? Are they encrypted with a key we control, and can they sit in an HSM or external key manager?
- Can files be deleted automatically after delivery, including copies in web portals, temporary folders and backups?
- What exactly is logged for each transfer, and how do we stream it to our SIEM without a custom parser?
- How would we detect a bulk download by a single account or a new administrator being created out of hours?
- For the hosted service: where is data stored, who at your company can access it, how quickly are emergency patches applied, and what are your breach notification timeframes?
- How do you handle partner onboarding and offboarding, including expiring keys and accounts that have not been used for 90 days?
- Which cryptographic modules do you use, and are they FIPS 140-3 validated if we need that? See the FIPS 140-3 and Cryptographic Module Validation Guide.
Red flags
- Silent patching: fixes shipped without advisories, CVEs or indicators of compromise leave self-hosted customers unable to judge exposure.
- The database or admin console on the internet in the vendor's reference architecture.
- Files kept indefinitely by default, with retention left as a customer task.
- Shared partner accounts or one service account used by all scripts, making revocation and attribution impossible.
- Logs only visible in the product's own console, or exported without file names, sizes or source addresses.
- Cloud storage connectors that need long-lived access keys typed into the configuration.
- No support for removing plain FTP and outdated algorithms, or a vendor that says partners "need" them without offering a migration path.
Proof of concept
- Pick representative flows: choose three to five real flows covering SFTP, one legacy protocol, a web portal user and an API or cloud connector, using test data only.
- Build the reference architecture: deploy with the gateway tier, admin interface on a management network and plain FTP disabled. Record every port that must be open.
- Run an external scan of the exposed tier with your usual vulnerability scanner and attack surface tooling, and review what is visible. See the Attack Surface Management Guide.
- Test credential handling: onboard a partner with an SSH key, set an expiry, rotate it, then offboard the partner and confirm the key and account are gone.
- Test retention: deliver files and confirm they are deleted from every location, including portal storage and temporary folders, after the configured period.
- Test monitoring: stream logs to your SIEM, then simulate a bulk download by a test account and an out-of-hours admin change, and confirm alerts fire.
- Test recovery and rollback: back up configuration and keys, rebuild a node from backup, and confirm partner flows resume. Keep the legacy route available until cutover is signed off for each flow.
Measure after go-live: track time from vendor advisory to patch applied, percentage of partner credentials with an owner and expiry, number of files held beyond their retention period, percentage of flows on SFTP, FTPS with TLS 1.2 or later, or HTTPS, and number of legacy FTP flows remaining. Review partner credentials that have not been used for 90 days each quarter and disable them unless the business owner confirms they are still needed, and rehearse the response to a vendor emergency advisory at least once a year so the patch, log review and credential rotation steps are practised before they are needed.
Framework mapping
| Framework | Reference | Relevance |
|---|---|---|
| NIST CSF 2.0 | PR.DS-01, PR.DS-02, PR.PS-02, ID.RA-01 | Protecting data at rest and in transit, maintaining software, identifying vulnerabilities |
| NIST SP 800-53 Rev. 5 | SC-8, SC-8(1), SC-13, SC-28 | Transmission confidentiality and integrity, cryptographic protection, information at rest |
| NIST SP 800-53 Rev. 5 | AU-2, AU-6, IA-5, CM-7, SI-2, RA-5 | Event logging and review, authenticator management, least functionality, flaw remediation, scanning |
| MITRE ATT&CK v20 | T1190, T1048, T1567 | Exploitation of the public-facing server and exfiltration of bulk data |
Standards and references
- CISA: Known Exploited Vulnerabilities Catalog
- IETF: RFC 4130 Applicability Statement 2 (AS2)
- IETF: RFC 4217 Securing FTP with TLS
- NIST: SP 800-53 Rev. 5 Security and Privacy Controls
- NIST: Cybersecurity Framework 2.0
- MITRE ATT&CK: T1190 Exploit Public-Facing Application
How NHI Mgmt Group can help
We provide independent requirements, RFP and evaluation support. Browse vendors in our products directory or contact us.
Related NHI Mgmt Group resources: Managed File Transfer (MFT) Security Guide · MOVEit Transfer Hardening Guide · Data Extortion Without Encryption Guide · Secure File Sharing Guide · Encryption Buyer's Guide