What is Fair Processing? 🤔
Enterspeed processes a lot of data in our Speed Layer. To ensure that every tenant and client is given equal processing time and resources, we implemented what we call Fair Processing. This means that no one customer will be able to hog resources and cause delays for others.
Most often, processing is very straightforward: One source entity is ingested, and a transformed view is ready for the Delivery API in a few seconds. No muss, no fuss.
But sometimes, the processing becomes complicated.
For instance, in the case of a schema deployment or a re-ingest of all source entities where the sheer volume of data might result in thousands of processing jobs that makes instant processing time near impossible.
In our Speed Layer, data processing is built as a FIFO queue (First In, First Out). As our layer is used to run multiple projects for multiple customers, we might run into incidents of backlogged processing queues that can affect processing times for other customers. Processing times are not always instantaneous and that can become even more complicated when processing thousands of processing jobs.
“The noisy neighbour problem” is a common issue that arises when one customer can affect others in a shared environment with shared resources.
By using our fair processing queue, we can regulate how much processing time is available for different classes of customers and different types of processing jobs. This ensures that every tenant receives an equal share of the available processing time and resources, making the processing queue fair for everyone.
In other words, fair processing allows us to remove the noisy neighbour problem so all customers will experience that ingesting a source entity results in a generated view a few seconds later.
No more over-processing
Another issue Enterspeed addresses in our fair processing queue is "over-processing".
Over-processing happens when updating one source entity results in a spike in processing jobs, resulting in the re-processing of too many source entities. By employing a job de-duplication feature, we can detect duplicate jobs within the same batch of processing jobs and reduce the overall number of jobs.
This saves both time and cost and is more sustainable.
So, the fair processing layer ensures that processing times are efficient and sustainable for all customers. It also eliminates the noisy neighbour problem, ensuring that each customer gets their fair share of processing resources without negatively impacting others.
Additionally, the introduction of job de-duplication has reduced processing jobs and resulted in significant savings in time, costs, and sustainability.
These improvements are part of Enterspeed's efforts to provide an intelligent caching solution that optimises performance, reduce costs, and minimises the environmental impact of processing.
20 years of experience with web technology and software engineering. Loves candy, cake, and coaching soccer.