An immutable stream of data that contains a time component. The data stream is immutable due to the fact that the data isn't meant to change. One example is a sensor that monitors heat of an industrial component. At a specific point in time, that component is X degrees. It will never be a different temperature. Logs could also be timeseries data as they happen at a specific point in time.
The name is a bit misleading as there are servers. Serverless basically means you do not have to manage servers. Serverless is best known for Function as a Service (FaaS) platforms. When you execute a function, you pay for the memory consumed and the length of time it takes your function to run and the cloud provider manages the rest.
Serverless is the next evolution of the cloud. Serverless frees the developer from having to worry about capacity and scaling. Hyprcubd is bringing the concept of serverless to timeseries databases.
One of the biggest hurdles of using a timeseries database is the operational complexity required. Timeseries data arrives at such a high velocity, traditional methods of ingestion quickly fail under even a modest load. There are many timeseries databases on the market today however operating them is not trivial. Hyprcubd was designed to handle this velocity and scales to meet peek demand.
Hyprcubd completely manages your data's lifespan from ingestion to deletion.
When your data is inserted into Hyprcubd, it is triple replicated to fault tolerant network storage devices. After a certain amount of time, your data is written to a highly durable object storage system with eleven 9s of durability.
Traditional DBMSs are designed for Create, Update, and Delete operations. To ensure data consistency, these databases must use transactions or locks to protect access to the same data by concurrent operations. This comes at a great cost to database complexity and storage design. Locking is slow especially in distributed systems. Storage for these systems is also row oriented on disk to provide fast lookup to the data being modified.
Hyprcubd is a purpose built timeseries database that reduces internal complexity by not requiring coordiated access to data. Data is also stored in a columar fashion to increase data locality, heavily increase processor cache performance, and utilize data compression techniques. All of these greatly increase the performance of Hyprcubd while reducing operational costs. We are able to pass these savings on to our customers.