Google Summer of Code

Google Summer of Code (g.co/gsoc) is Google's mentorship program for bringing new contributors into open source communities. It's happening again for the 19th year in 2023! Over 18,000 developers from 112 countries have participated.

Google Summer of Code is a unique program where new contributors to open source, ages 18 and over, are paired with a mentor to introduce them to the open source community and provide guidance while they work on a real world open source project over the summer. Projects cover a wide range of fields including: Cloud, Operating Systems, Graphics, Medicine, Programming Languages, Robotics, Science, Security and many more. GSoC Contributors do earn a stipend to work on their medium (~175 hour) or large (~350 hour) projects. This is not an internship but provides an invaluable experience and allows you to be part of an amazing community!

GSoC is a highly competitive program, so don't wait to the last minute to prepare! GSoC Contributors should reach out to the organizations that interest them once orgs are announced on May 4, 2023. Potential GSoC Contributors can apply at g.co/gsoc from March 20 - April 4, 2023.

See the list of projects we have available for GSoC contributors and learn how get started with contributions.

Disk Fragmentation Simulator

Mentor name(s): Adam Kupczyk

Mentor email(s): akupczyk@ibm.com

Difficulty: Intermediate

Project Hours: 350

Skills needed: C++, Random Distributions, Data Presentation

Subcomponent of Ceph: BlueStore

Description of project: BlueStore is a component of Ceph responsible for storing objects. Object is a data container, like a file with extra capabilities. To assign an object some free disk region BlueStore uses a disk allocator component. When BlueStore was created it came with a simple disk allocator. That worked pretty well in testing and even on customer deployments.

Customer workloads sometimes have complicated write patterns that expose allocator deficiencies. In most cases the observed negative effect is object fragmentation. Fragmentation is not a critical issue, an object that is distributed on the disk to multiple fragments is still valid, but accessing it drags performance.

Long term goal is to augment BlueStore with fragmentation control / combat features. We understand that implementing algorithms is costly and we want to evaluate quality of certain concepts before we start implementation. We expect to see in each algorithm some emergent properties, and we need to squeeze months of customer operation into minutes.

We need a disk fragmentation simulator.

The simulator will provide an environment to test prospective allocators. It will apply data-trashing workloads. Workloads will reflect the fact that customer patterns usually change from time to time. Object access distributions should periodically change; peak, hot objects set, r/w access size. Extra random events that are specific to Ceph OSD (placement group add, placement group remove) can be added later.

The simulator must measure the quality of object fragmentation. It would be good to have 2 types of measured fragmentation score: static per-object and dynamic related to customer workload currently simulated. A great extra would be an ability to represent fragmentation as some kind of diagram, use color as presentation aid, and show the evolution of fragmentation in time.

The simulator platform should be written in C++ but must incorporate the existing Allocator interface. The platform should expose C++ interface toward "defragmentation algorithm".

A contribution in the form of a new allocator proposal or defragmentation strategy is welcomed but secondary. The goal is to evaluate allocators as we want to root out or at least document behaviour of a long running system.

Standup/weekly call mentee could attend?: Core standup & Bluestore Upkeep & Evolution

Steps to evaluate an applicant for the project:

Applicant should email the mentor (akupczyk@ibm.com) for detailed instructions on the projects application process.

In a process of assessing applicant, one will:

Gotta Catch 'Em All

Mentor name(s): Yuval Lifshitz

Mentor email(s): ylifshit@ibm.com

Difficulty: Intermediate

Project Hours: 175

Skills needed: C++

Subcomponent of Ceph: RGW

Description of project: Use the Coverity static snalysis tool to cleanup issues from the RGW code.

Applicants can get a more detailed overview of the project here

Standup/weekly call mentee could attend?: RGW standup

Steps to evaluate an applicant for the project:

Applicants should build ceph, analyze and fix a simple coverity issue.

To learn how to get started visit this link