I am somewhat surprised to see this story on a lack of a backup. Apparently the NYPD has a system where they track property and evidence. It's a big system, one that cost $25.5 million. This came up as a part of an investigation into cash forfeitures, when the department's lawyer said that requests for information couldn't be provided because the system has no backup. At least, that's the first part fo the story.

This gets confusing because the system isn't designed to provide an audit report from the database, and apparently they can't query the database. It seems there are replicated backups to other systems, but the entire thing is live, so there's a danger that any misuse could disrupt the system and crash everything. The lawyers are even claiming web scraping might cause issues.

In a followup article, the vendor's description of the system mentions a DB2 database, but the NYPD lawyers have said the system isn't based on an DB2 database. It's possible the database was replaced, but stil, there has to be some database, right? Relational, NoSQL, flat files. There's some storage mechanism.

Ultimately, this seems like a legal trick, not based on any technical understandings. Webscraping would be similar to any other querying of the system. Accessing the underlying database shouldn't be hard, and certainly a backup must be maintained somewhere. If those aren't possible, then clearly there are multiple nodes. Take one offline and query that node. Most of us that work with applications know that the system isn't this fragile. While there may not be offsite tape or disk backups, the whole story seems fishy.

Most of us that work with databases know what backups are critical. In fact, I'd say that the most important thing we do is ensure there are backups avaliable in the event that we need to restore. Systems are fragile, and any particular piece can crash. Hardware fails and needs to be replaced. Power fluctuations or temperature changes can cause corruption. We always need backups.

At some point I'm sure this will be shown to be a bit of a scam and thre will be some backup taken of the system. This is regardless of whether there have been any backups taken before this point. It is possible that the IT staff or the maintenance vendor isn't doing their job, but having a lawyer stall or hide this fact might work for awhile.

Most of us won't have those circumstances. We might be able to lie about not taking backups, or losing some and be fine, but we might also be dismissed for cause. Don't put yourself in a bad situation. Make backups, make sure you can restore them, and make sure that system runs often. Let someone else decide what to tell clients, customers or judges if there are issues, but make sure you've performed your job as a data professional.