Notice
|Available in VPC.
To enhance service quality, HEaaN Homomorphic Analytics will be temporarily unavailable starting from September 5, 2024. We will inform you as soon as a specific date is confirmed. We apologize for any inconvenience.
Analyze and calculate encrypted data in cloud with HEaaN
Fully homomorphic encryption (FHE) is a form of next-generation encryption that allows users to compute encrypted data without decrypting it. CryptoLab's HEaaN available on NAVER Cloud Platform is the first homomorphic encryption library that allows users to compute real numbers. HEaaN is widely recognized in academia and has demonstrated its excellence in international competitions with 14 associated patented technologies.
The cloud protects your data through various security technologies. With homomorphic encryption, you can directly combine, compute, and analyze encrypted data with no restrictions. Thus, data remains encrypted throughout the process of being shared, saved, and analyzed, guaranteeing that no data is exposed in its unencrypted form.
You can derive valuable insights by using statistical analyses and machine learning on strongly encrypted data without compromising security.
Finance, healthcare, and public service providers are subject to regulations regarding personal, sensitive information. To comply with these regulations, data need to be de-identified for statistical analyses. Homomorphic encryption is the solution to guarantee data security while complying with such privacy regulations for the service providers in the finance, health, and public sectors.
NAVER Cloud Platform allows you to easily leverage a variety of data analysis tools involving homomorphic encryption on the web console without extensive understanding of the technology. Plus, you can use Private Instance to create a self-contained analysis environment to run data analysis and view results instantly on the Jupyter Notebook.
Data of National Pension Service and KCB (Korea Credit Bureau) has been combined and analyzed within the Innovative Financial Services' analysis frame using homomorphic encryption technology.
You can process encrypted data, such as in extracting, copying, moving, and deleting them.
Plus, you can run arithmetic operations such as addition, subtraction, multiplication, and division as well as run descriptive analysis such as computing the mean, variation, standard deviation, standard error, and coefficient of variation.
Data are normalized for training and model fitting using logistic regression.
You can refine your logistic regression models by evaluating machine learning models.
HOMEROS is a Python library available on NAVER Cloud Platform that supports computational operations and analysis using encrypted data empowered by homomorphic encryption. Other than computations provided by HEaaN, high-level functions required for data analysis and machine learning are available so that data analysts and developers can access them with ease.
Homomorphic encryption keys are essential to data sovereignty.
You can create homomorphic encryption keys and encrypt data by using an agent that can run in your environment.
You can perform computations using encrypted data in the cloud and save them to a local environment to decrypt the computation results.
Private Instance is a server instance that runs in a virtual private cloud (VPC) environment.
You can use the Jupyter Notebook packaged with HOMEROS SDK to write code and check the progress in a simple, easy manner.
In addition, it makes development much easier by allowing you to develop services using homomorphic encryption, which can be integrated with existing services as a microservice.
Classification | Billing standard | Pricing |
---|---|---|
Homomorphic encryption compute time | Hour(s) | - |
(VAT Excluded)
Compute time includes creating homomorphic encryption keys as well as encrypting and decrypting data in the cloud.
Compute time is billed hourly, not by minutes or seconds.
Homomorphic encryption keys, data for computation, and computation result data are stored in Object Storage. Additional fees for using Object Storage apply.
Downloading homomorphic encryption keys and data incurs additional fees associated with outbound network traffic.
Type | GPU | vCPU | Memory | Disk | Billing standard | Usage fee |
---|---|---|---|---|---|---|
CPU Intensive | - | 8 units | 16GB | SSD 50GB | Hour(s) | - |
CPU Intensive | - | 16 units | 32GB | SSD 50GB | Hour(s) | - |
CPU Intensive | - | 32 units | 64GB | SSD 50GB | Hour(s) | - |
GPU | 1 units | 8 units | 90GB | SSD 50GB | Hour(s) | - |
GPU | 2 units | 16 units | 180GB | SSD 50GB | Hour(s) | - |
GPU | 4 units | 32 units | 360GB | SSD 50GB | Hour(s) | - |
(VAT Excluded)
Private Instance supports only a set of server types and specifications to ensure performance (Only V100 is supported for GPU).
Private Instance is billed as HEaaN Private Instance in your bill.
Additional fees for network traffic apply. For more information, see the network fees of Server.
Create a storage space to store keys and data encrypted in HEaaN. For more information on pricing, see Block Storage for pricing information.
Pricing scenario
Homomorphic encryption compute time: 221 hours 20 minutes
Creating homomorphic encryption keys: 5 full types (14 GB per key)
Computing and result data: 20 units (assuming the size of encrypted data is 100 MB)
Estimated fee
Compute time fee: 222 hours x - = -
Object Storage cost
Key data: 14GB x 5 keys = 70GB
Computing and result data: 100MB x 20 units = 2GB
Total: 72GB x - = -
* If you download data, then additional fees for outbound traffic will apply.
Usage scenario
Instance type: [CPU Intensive] vCPU 8 units, memory 16GB, disk 50GB 1 unit
Usage hour(s): 100 hours
Estimate
Usage fee: 100hours x - = -