Edge or Cloud? Coupling the benefits of both computing concepts for Data Science

Standard

Shall I deploy my application on the edge or in the cloud? This is a question that a Provider for a SaaS application will ask him-/herself more often than not. After exploring and cleaning your data, deriving a use-case for an analytical service, and finally implementing it, you will need to take the decision on where to deploy your application to offer the full functionalities and benefits to the end user.

The benefits of edge computing include low latency, cost efficiency in terms of data migration and energy and independence from stable internet connections. On the other hand, cloud computing offers more computational power, resiliency, flexible and cost-efficient hardware. Both areas have their raison d'être and offer various advantages for deploying analytical services and software.

In a perfect world, we would also like to combine these two concepts.

Two examples are:

  • Machine Learning: Train the model in the cloud and use it on the edge
  • Run an application on the edge, but move to the cloud in case it needs more resources on demand

Are there any tools to help with the decision? Do we even have to decide, or can this process be automated?

The good news is PLEDGER will offer toolkits to help you to get the best from both computing concepts.

To help to find the appropriate infrastructure PLEDGER benchmarks the available infrastructure, profiles the application, and creates a profile to facilitate the decision for the system taking also preferences into account. For the application provider it is important to provide meaningful metrics to measure the Quality of Service (QoS), which will be monitored by the system. In case QoS decreases,

PLEDGER detects this violation, finds an alternative infrastructure, and issues the migration of the application.

PLEDGER will not only help to decide between edge and cloud – it will even automate the process.

logo_inverse

is loading the page...