REX: A Development Platform and Online Learning Approach for Runtime Emergent Software Systems


Authors: 
Barry Porter†, Matthew Grieves†, Roberto Rodrigues Filho† and David Leslie‡
†School of Computing and Communications; ‡Department of Mathematics and Statistics Lancaster University, UK

Overview: 
It's a very interesting paper. This paper gives a new idea for developing software: components and runtime adaptor, and it divides the runtime adaptor into 3 layers. First, the paper designs a new "glue" language named Dana to describe component interface and low level relationships. Then it proposes the "Perception, Assembly and Learning" layer (PAL) for: connecting components (assembly) to a full software system, collecting component's metrics/performance (perception) and provide interface for the learning layer. The highest learning layer will analyze the provided metrics and given action feedback for dynamic tuning.

Main points in this paper: 
REX: Dana, PAL and Learning
Dana:
- Describe components' interfaces and dependence relationship
- A protocol for seamless runtime adaptation (component to component)
- Object functionality: private state instantiation and pass references as parameters
PAL:
- SetMain for deciding the major software service, assembly the entire software system
- "Recorder" interface for monitoring components behavior
Learning:
- Baseline algorithm is multi-armed bandit problem and Thompson sampling
- Linear regression model for differentiating independent factors (components)
- Extra term for handling runtime environment in learning equation

Strength: 
- A new direction for software design (pretty like hardware design: IP and system modeling process).
- It demonstrates good design in: system assembling, runtime learning and adaptation.
- It considers several variances between ideal learning algorithm and real software system like: shared components metrics and runtime reward changes.

Concerns: 
- It's not clear the design handling data movement or other side effects leaded by reconfiguration/re-assembly. For sophisticated storage component, does REX work well in avoiding/mitigating migration overhead? Or for policy changing like cache replacement, how to transfer the old policy attributes to the new attributes (side effect)? e.g. if cache policy moves from LRU to LFU, the timestamp is useless and frequency information is needed.
- For adaptation protocol, REX provides adaptation at component level but misses for the entire software level, i.e. the adaptation order for many-components (which one adapts first?).
- Developers need to provide the complete components for REX to choose.

Own opinion: 
This paper gives new idea about how to develop software system: more structured and runtime learning based reconfiguration. Even though some details not clear, but I pretty like this direction.


Blogger Minjun Wu, 04/03/2019

Comments

  1. jon_weissmanApril 4, 2019 at 5:54 AM

    You are correct that certain adaptations may require very high overhead. The system might decide to avoid an adaptation that was too expensive if this cost was in the model. You're right that some adaptations might be very complex: it seems that the approach might require 'safe' points for adaptation. You raise another interesting point about component groups being the granularity of adaptation.

    ReplyDelete

Post a Comment

Popular posts from this blog

A Machine Learning Approach to Live Migration Modeling

Authors:  Changyeon Jo, Youngsu Cho, Bernhard Egger Seoul National University Summary:  This paper proposed a machine learning approach to predict the system metrics for different VM migration algorithms. These system metrics are affected by migration algorithm, host machine status and user applications and feasible for machine learning approach. The prediction model takes 21 system/application input features (including which algorithm), and generates 6 metrics as the output. As a result, it can accurately estimate each metrics under different application, VM and algorithms. One of the application, the paper provide a semi-automated algorithm selection based on this machine learning model and user-provided SLA (service level agreements) constraints. Questions about Machine Learning:  Why machine learning is applied: The combination of algorithm, VM status and current running applications status, is a huge searching space. Machine learning features: which algor...
Read more

StormDroid: A Streaminglized Machine Learning-Based System for Detecting Android Malware

Introduction: The paper approaches the problem of Malware Detection using machine learning approach. In this paper, the authors propose considering new type of features along with two traditionally used set of features for developing techniques for malware detection. Along with this proposal, the paper makes other contributions as well. The authors provide a new framework which is able to process information from large data sets as streams, and efficiently perform malware detection. Motivation: Android is an open platform for development. It has certainly made easy for everyone to develop and publish apps.  The number of apps published on Google store is in millions. However, it has also made easier for hackers to reach out to a larger population for malicious attacks. Additionally, with surge in the number of attacks on devices, more data is available for analysis which can be used by the machine learning models to identify malicious apps. This is motivation to use machine ...
Read more

Coordinated Management of Multiple Interacting Resources in Chip Multiprocessors: A Machine Learning Approach

Authors: Ramazan Bitirgen, Engin Ipek, Jose F. Martinez Motivation: Sharing system resources efficiently on chip multiprocessors (CMPs) platforms is the key to obtain high utilization and enforce system-level performance objectives. Unrestricted sharing of microarchitectural resources will result in destructive interference and unacceptable performance. However, the works focus on coordinated management of multiple interacting resources on CMPs are very scarce , even though several proposals focusing on addressing the management of a single microarchitectural resource exist. Aiming to fill this gap, this paper proposes a framework with the machine learning technology to manage multiple shared CMP resources in a coordinated fashion to enforce higher-level performance objectives. This novel mechanism consists of per-application hardware performance models and a global resource manager. Positive Points: 1. The global resource manager is able to aggregate each app...
Read more