Serverless computing at the edge for the Internet of Things
| dc.contributor | Aalto-yliopisto | fi |
| dc.contributor | Aalto University | en |
| dc.contributor.advisor | Virtanen, Valtteri | |
| dc.contributor.author | Tudlik, Zoltán | |
| dc.contributor.school | Perustieteiden korkeakoulu | fi |
| dc.contributor.supervisor | Di Francesco, Mario | |
| dc.date.accessioned | 2019-06-26T11:06:07Z | |
| dc.date.available | 2019-06-26T11:06:07Z | |
| dc.date.issued | 2019-06-17 | |
| dc.description.abstract | The number of Internet of Things (IoT) devices, such as sensors, cameras, and actuators, is increasing exponentially, along with the amount of data generated by them. The traditional approach to processing and storing IoT data is by using cloud data centers that are often geographically far away from the source. However, the underlying network is unable to endure the increased load and comply with strict low-latency requirements of several IoT applications. This thesis studies the proposed solution, edge computing, that moves computation to the proximity of IoT devices. First, the thesis surveys IoT applications and their quality of service requirements that can potentially be satisfied by using edge computing. Second, the author develops an image recognition software to benchmark two edge computing platforms, Amazon Web Services (AWS) Green- grass and Azure IoT Edge and a cloud computing platform, AWS IoT. The first experiment uses this benchmark to compare the round-trip latency provided by AWS Greengrass and Azure IoT Edge and to estimate the incurring monthly operation costs at a scale. Third, to better understand the benefits gained by using edge computing platforms, the second experiment uses the same application to compare the network and processing latencies and operation costs provided by AWS Greengrass and AWS IoT. The survey of IoT applications shows that without the advancement of 5G technology, edge computing is currently the only way to meet latency requirements if it is possible at all. Experimental results show that AWS Greengrass provides the lowest overall latency of the three platforms, while Azure IoT Edge is the cheapest one if communication with the cloud is not necessary. Using either edge computing platforms imply compelling cost reduction compared to AWS IoT. | en |
| dc.format.extent | 61 | |
| dc.identifier.uri | https://aaltodoc.aalto.fi/handle/123456789/39147 | |
| dc.identifier.urn | URN:NBN:fi:aalto-201906264212 | |
| dc.language.iso | en | en |
| dc.programme | Master's Programme in ICT Innovation | fi |
| dc.programme.major | Software and Service Architectures | fi |
| dc.programme.mcode | SCI3082 | fi |
| dc.subject.keyword | serverless | en |
| dc.subject.keyword | edge computing | en |
| dc.subject.keyword | cloud computing | en |
| dc.subject.keyword | Internet of Things | en |
| dc.title | Serverless computing at the edge for the Internet of Things | en |
| dc.type | G2 Pro gradu, diplomityö | fi |
| dc.type.ontasot | Master's thesis | en |
| dc.type.ontasot | Diplomityö | fi |
| local.aalto.electroniconly | yes | |
| local.aalto.openaccess | no |
