What are you looking for ?
Infinidat
Conference_Concepts

Start-Up’s Profile: Cerabyte

In long-term storage using ceramic nanolayers for cold data

Name
Cerabyte (Ceramic Data Solutions Holding GmbH)

HQs
Munich, Germany, currently Balanstraße 71; from October 1 Rundfunkplatz 2

Date founded
January 6, 2022 (Holding) owns Ceramic Data Solutions (2012) and Ceramicro (2021)

Financial funding
Public grants (2020) €1,2 million, seed round (2022) €6,5 million by 3 founders, Swiss VC Aeony (company owned by the founders, as investment vehicle for minor investors)

Revenue
No one yet

Main executives
They are specialized in material technology, long-term information storage, laser physics, engineering and IT.
Pflaum CerabyteChristian Pflaum, co-founder and CEO

 

 

Number of employees
10 FTE at Cerabyte, 20 FTE for Cerabyte at project partners

Technology
Permanent storage without energy and frequent replacement of media, using ceramic-based data carrier and fs-laser matrix for writing and high speed microscope cameras for reading, we are introducing a new storage technology using inorganic nano layers using ceramics that are 50-100 atoms thick to store information protected vs. most storage media threats. Data can be written and read using laser or particle beam structuring the information in data matrices, in similar structures as the widely used QR-codes. The technology roadmap is projected to scale from 100nm to 3nm bit sizes, corresponding to areal densities of GB/cm2 to TB/cm2 class. The use of high-resolution microscopic imaging techniques for optical read or electron beam microscopy, for structures below the optic diffraction limit, enables the reliable retrieval of the data from the media. The volumetric data density mainly scales with the thickness of the substrate, which can be 100-300µm thick sheets or 5µm thick tapes coated with 10nm thick ceramic coating. Particle beam writing on tape could achieve TB/mm3 class exceeding the storage density of all commercially available storage solutions today by an order of magnitude. Laser beam matrices will enable datacenter rack storage densities scaling from initial 10PB systems to 100PB using CeraMemory cartridges (2025-2030) and 1EB using CeraTape (2030-2035).

Product description
New tier for long-term storage, for cold data, which in fact are 70+% of data in datacenters, as good as never retrieved but stored for decades without intention to be ever deleted, and permanently consuming energy and resources. The firm uses ceramic nanolayers creating the most reliable storage ever. Data is save in a temperature range of -273°C (-460°F) to 300°C (570°F) and even in corrosive or acidic atmosphere. Also radiation and an EMP can’t destroy information stored on Cerabyte. Form factor is 19-inch rack-size.

Cerabyte media technology is leveraging existing production capacities of hundreds of millions of square foot of ultra-thin display glass enabling a media cost scaling path to less than $1/TB by 2030.

Released date
2024

Price range
Eventually 75% less cost than long-term storage on HDD

Roadmap
Demo system end of 2023, pilot projects 1Q24-2Q24, storage systems on-prem 4Q24, full RW systems on-prem 2025

Partners
Manufacturer of thin glass, laser industry, PVD industry, robotic storage manufacturer

Distributors
Partners from Saas, later licensing to hyperscalers

Main customers
Data centers, hyperscalers, corporations and entities with archival data on-prem

Applications
Infinite and low cost storage, highly resilient to influences that destroy conventional memory media (water, fire, corrosion, magnetic pulses, etc.)

Target market
Hyperscalers and data centers who need to securely keep customer’s data and spend comparably enormous expenses on data with little value.

Competitors
MS Silica, Dots (Group 47), Folio Photonics, firms in DNA storage

Read also :

Comments

Kunze CerabyteHere are more information given by Martin Kunze, founder and CMO:

 

 

The entire system eventually will consist of a writing unit (19" rack footprint) with the femtosecond-laser source that can feed multiple writing heads, (depending on miniaturization: 20 or more) with writing speed of GB/s each.

The data carrier itself is a square thin glass sheet (similar to Corning's Gorilla Glass) with a thin dark ceramic coating. The writing head uses a laser matrix of up to 2 million bits per shot, generated with the help of a digital micromirror device (DMD) with repetition rates of kHz, hence the high writing speed.

The glass sheets are stored in cartridges, which enable random access and short time to first byte of somes. Cartridges are kept in libraries in 19" rack footprint. Each library rack has at least one reading head (depending on the projected retrieval rate) with GB/s. A horizontal transport system delivers the written, full cartridges to the library racks. When a row of racks is fully written, die writing unit is moved to the next row and starts writing 24/7 again ...

The data capacity will be 10PB per rack with a roadmap to hundreds of PB/rack.

As you may understand it is not a consumer product. It is a storage solution for the vast amount of cold data in datacenters.
 
This is how it is projected.
 
The company is currently assembling the demo-system, which has of course not at all the above-described performance. However, our principle is to solely utilize (and partner with their supplier) materials and technologies, that are already widely used - to enable later fast scale up. Therefore, we only use components off the shelf for the demo system. The novelty is to combine them in some parts unconventionally. This allows us to build a demo system in shortest time to prove the concept is working.

Several components need adoptions to scale up performance, for that we are collecting funds in the A-round next year.

Example: for reading we use one of the fastest image sensors, which is - as it is standard today - a color sensor (RGGB) with a 16bit color-depth. For our purpose, we need a monochrome sensor with maximum 3bit greyscale. This would increase the resolution by 4x and speed up the image processing by 1 or 2 orders of magnitude and together increase the reading speed to the projected GB/s. Developing a monochrome sensor with low bit-depth is no rocket science and credibly arguable.
 
We created a handful of short videos for our website, that throw light on Cerabyte from different perspectives: TCO, reducing the digital age's carbon footprint, resilience, look inside, Cerabyte for archival data, Digital data as durable as hieroglyphs. A shortcut to the videos is here.
 
The idea to use ceramic as data carrier stems from a project I started about 10 years ago where images and text are printed with ceramic colors on high-quality bathroom tiles, stored deep in the oldest saltmine in Austria in order to pass on a snapshot of our era - and to point to the problem in the first place of how to preserve information at all on the long run. The MOM project was and still is widely covered in media and it helped me to find the respective experts and partners to eventually create a storage system that keeps data literally forever accessible, while also solving the problem of high costs (economically and ecologically) of storage.
 
In contrast to all other emerging technologies for long-term storage, Cerabyte origins from the purpose, while all others come from technology: some physicist, microbiologist, or chemist discovered a phenomenon or an effect and came up with the idea to use this for storage. We in contrary addressed the problem from the pain-side - utilizing a material with a track record of 5.000 years as data carrier ...

Articles_bottom
AIC
ATTO
OPEN-E
RAIDON