In partnership with

PreSeed Now has been running for over two years now, and it’s funny how certain names keep cropping up over time.

We’ve featured two completely different startups called Blend in the past, for example. And today is the third time we’ve spoken to someone called Ed French.

This Ed French isn’t the same one behind adtech startup Contxt Media, but he is the same one who is involved with carbon capture startup Alkama.

Today’s startup, Umbra Systems, is doing something very interesting with cloud data. Read on to find out what it is…

– Martin

Intercom for Startups

Join Intercom’s Early Stage Program to receive a 90% discount.

Get a direct line to your customers. Try the only complete AI-first customer service solution.

Apply now

Umbra Systems hides sensitive cloud data in the shadows

Cloud security is a perennial problem. Even if you secure your data with the best available encryption, you don’t know when that encryption might be cracked.

And even if it is secure, there’s the slight chance your cloud provider might accidentally delete all your data and backups, as Google did recently to a big Australian pension fund. Yikes.

Umbra Systems has a fresh take on cloud security and data management that aims to make these risks a thing of the past.

How it works

The Manchester-based startup is doing something unusual to better secure cloud data - it’s saving it in the shadows.

Instead of encrypting the data in a conventional way, Umbra makes ‘shadows’ - files in which the original data has been completely replaced by a much smaller collection of other bytes using clever mathematics. The same maths can then be used to reassemble the data into its original form. 

The data is broken up into seven or more shadows, which are then encrypted and stored across multiple cloud providers. Interestingly, not all the shadows are needed to restore an original file. Any four of the seven shadows is required to reassemble the full data.

The technology is based on the secret sharing algorithm created by Adi Shamir in 1979, but with a modern twist to make it work practically in a cloud storage context.

“If you remember back in school when you did simultaneous equations, we've got four variables we care about, and we've made seven equations. If you've got four variables you needed to solve for, you couldn't solve the equations till you've got four. And it's the same principle, but on steroids, that's going on here,” explains founder Ed French.

“Each of the shadows is a quarter of the size of the original and doesn't contain any of the bytes of information that were in the original.”

This means a hacker would need to access multiple platforms and unlock four of the ‘shadows’ to recreate the original data.

“We need only four of the shadows, and it doesn't matter which four. We don't rely on any of the cloud services maintaining uptime, or indeed keeping the data safe. So if one of them deleted everything, it doesn't actually affect everything,” says French.

And while the security world is looking to solutions like quantum encryption to protect files in the future, French argues his system is superior.

“Until you've managed to hack four separate repositories, you haven't got enough data there to start to decrypt it. [With fewer than four shadows] you've only got information that tells you some things in aggregate about the whole thing, but not enough for you to be able to derive what the information was.”

Because the shadows are much smaller than traditional backups, there’s the potential to save money on cloud storage and pay for Umbra’s service with the savings, French says.

The story so far

French has a background of working with, and investing in, early-stage startups. We last encountered him as executive chairman of carbon capture startup Alkama a couple of years ago. 

He says the idea for Umbra Systems came to him while talking with the founder of a startup serving the legal profession, a typically cloud-averse market because of the potential security risks of remote storage.

“People want the flexibility of cloud-based services. They want to make it simpler to build things rapidly. They want scalability. All those things are incredibly useful. But at the same time, there's a movement back away from the cloud now to on-prem, because of all these other issues that come up,” says French.

“I realised that [the Umbra Systems approach] was potentially an opportunity to make the cloud the safest place to store the data.”

French has worked with thestartupfactory.tech to build Umbra Systems’ technology, and the startup’s CTO has come from that company. French says the tech is now up and running and will be put to the test by a large telecoms company in their labs this month.

The startup is also talking with a number of potential partners and customers in fields such as cybersecurity, who handle sensitive data.

And as a test of its technology, Umbra plans to publicly release a shadow of data from an ebook in Project Gutenberg’s public domain collection.

French says it’s nearly impossible to figure out which book it is. But given the limited amount of source data (Project Gutenberg’s library of 70,000+ ebooks), it might be possible to narrow down the search to a handful of titles if someone could figure out something about the book in question from the data reconstructed from its shadow… if they can reconstruct it in the first place.

“We will offer a job to whoever manages to figure out which book we have published,” says French.

Subscribe to Premium to read the rest.