1,463 Leaked Records and the Quiet Case for Permanent Storage

On September 29, the Tokyo-based ticketing platform eplus disclosed that attackers had breached the system managing refund data for its digital ticket service, Smart Ticket. The breach exposed 1,463 users' personal information, and for 751 of them, the leak included bank account numbers and home addresses. Not passwords. Not email addresses. The exact details you need to drain an account or forge an identity.

That specificity matters. We tend to measure breaches by volume—millions of records, billions of credentials. But the damage scales with the sensitivity of the data, not the count. A thousand bank account numbers with matching names and addresses is a high-value target for fraud, and it exists because a live, internet-connected system was doing its job: processing refunds.

The Refund Problem Is a Storage Problem

Refund workflows are inherently awkward for security. They require a company to collect and retain bank details, verify them, and keep them accessible for a defined period—usually months, sometimes years. That data sits in a database that must be online, queryable, and connected to other systems. Every one of those requirements is a potential attack surface.

The eplus incident shows the pattern. The compromised system was not the ticket sales platform itself; it was the refund information manager. A specialized, lower-profile subsystem that still held highly sensitive financial data. Attackers understand this asymmetry. They target the ancillary systems where security budgets are thinner and monitoring is weaker.

Traditional cloud storage does not solve this. It replicates data across regions, encrypts it at rest, and gives you access controls. But the data remains live. It can be queried, modified, or exfiltrated by anyone who compromises the right credentials. The 1,463 records were not stolen from a backup vault; they were pulled from a working system.

What Permanent Storage Changes

There is a different model for data that needs to be kept but not constantly accessed: immutable, decentralized storage. On Arweave, for example, files are written once and paid for with a one-time fee that covers storage costs in perpetuity. The data is not held on a single company's server. It is distributed across a network of independent nodes.

For refund records, the implications are structural. Instead of a live database that must be defended continuously, you get an encrypted archive. Client-side AES-256-GCM encryption means the files are locked before they leave the user's device. The private key stays with the data owner. Even the storage platform cannot read the contents.

This is not a theoretical improvement. It changes the economics of a breach. If an attacker compromises the storage layer, they find encrypted blobs. If they compromise the user's device, they still need the private key. The attack surface shrinks from a constantly accessible database to a single, user-controlled secret.

The Economics of Keeping Data Forever

Subscription cloud storage is built on a recurring revenue model. You pay monthly or annually, and if you stop paying, your data is at risk. That model works for active workloads—shared documents, running applications, collaborative files. It is a poor fit for records that must be preserved for compliance or dispute resolution but are rarely accessed.

A one-time fee for permanent storage flips that equation. For a company holding refund records for seven years, the cost is predictable and front-loaded. For an individual archiving a signed contract or a family photo collection, it removes the anxiety of an expiring subscription.

This is the niche Long Drive occupies: permanent storage on Arweave, operated by Supercomputing International in Hong Kong, with optional client-side encryption and a free tier that covers files under 100 KB. It is not a replacement for a live database. It is a place for the data you need to keep but do not need to query every second.

The Real Question After a Breach

The eplus breach will be investigated, patched, and eventually forgotten. But the underlying issue will remain: we store sensitive data in systems designed for access, not for preservation. We treat every record as if it needs to be instantly available, and we pay for that availability with a larger attack surface.

The 751 users whose bank details leaked did not need those details to be live. They needed them to be accurate, retrievable when a refund was due, and then locked away. That is a storage problem, not a database problem. Until we separate the two, the next breach is already in someone's refund queue.