Patrick Buchner,
"From the Cloud to the Edge: An Infrastructure for Cloud & Edge Computing"
, 8-2019
Original Titel:
From the Cloud to the Edge: An Infrastructure for Cloud & Edge Computing
Sprache des Titels:
Englisch
Original Kurzfassung:
Providing services via a central cloud is an established method of building a scalable customer infrastructure using provider resources. The service quality of this centralised approach depends on the network between the customer and the central cloud. Despite this high importance, a company that is hosting services on a cloud has no in?uence on the quality of a network path.
A multi-cloud is presented as a federation of multiple, independent clouds which can be used to increase the number of data centres and vendors where services are provided. Utilising a multi-cloud instead of a single cloud can improve service quality in terms of availability or performance because a customer is not dependent on a single cloud provider. In addition, multiple providers may compensate the missing presence of a data centre from an other cloud provider to extend regional coverage and to reduce latency.
Edge Computing is presented as a general concept of locating services nearer to the user than a central cloud would be capable of. The extensions of the cloud, called ?Cloudlets? serve users in their area on behalf of the central cloud with lower latency. This is especially important in areas where the next cloud data centre is too far away to provide a reasonable performance.
Using these two generic approaches, a concrete infrastructure based on the requirements has been implemented with an industry partner from the media industry. This infrastructure is built similar to a hierarchical network, using a multi-cloud as redundant core and various edge installations as an access layer.