search

samvera-labs / samvera-connect

Samvera Connect Program
https://wiki.duraspace.org/display/samvera/Samvera+Connect+2019
Apache License 2.0
1 stars 0 forks source link

Fedora 4 performance: current status and future works #118

Closed jcoyne closed 6 years ago

jcoyne commented 7 years ago

Fedora 4 performance: current status and future works

Suggested by: Yinlin Chen Presenter: Yinlin Chen Format: Presentation Audience: All

jhriv commented 7 years ago

+1

kestlund commented 7 years ago

Combine with #138 from @awead

yinlinchen commented 7 years ago

Hi,

I am happy to co-present with Adam Wead. Please see below for my part of the content in this presentation. Thanks!

== Title: Samvera and Fedora 4 performance

Presenter(s): Yinlin Chen Affiliation(s): University Libraries, Virginia Tech Audience: All

Synopsis: In this presentation, we present Fedora 4’s performance in different use cases. These use cases are the most general use cases that were developed by the Fedora 4 performance team. We compare the performance between different versions of Fedora 4 and demonstrate the improvement between these versions.

We will also describe the procedures of these performance experiments and demonstrate the software we use (Ansible, AWS, and JMeter). Our performance experiments are conducted in the AWS ecosystem, and we use an R graph tool to visualize the JMeter result. With the AWS grant, we can do many different performance experiments in many scales. Moreover, it is easy to repeat and verify the performance results using AWS.

The Fedora 4 performance team wants to know more about the Samvera community’s interests related to the Fedora 4 performance. We plan to gather more use cases from the Samvera community, including middleware such as Solr, Camel, and Fuseki, etc. We can investigate, explore issues, and continue enhancing the future of Fedora 4.

awead commented 7 years ago

@kestlund Title: Performance Metrics with Hyrax and Valkyrie Audience: Developers and Managers

This presentation compares the performance metrics from a Hyrax-based application and a Valkyrie one. It will look at how the two perform with large collections of many thousands of unordered and ordered works. Comparisons will be made across a number of different factors including metadata-only resources, resources with metadata and binary data, persistence on disk, database, and Fedora, as well as using PCDM-based models and non-PCDM models.

Each platform offers various advantages and disadvantages. Conclusions drawn from this work hope to point at some of the reasons why scalable performance is problematic in the universe of Samvera applications.

These tests were performed as Penn State has explored which platform to use for its new cultural heritage object repository project.