OpenRASE: Service Function Chain Emulation

Theviyanthan Krishnamohan, Paul Harvey · 2025

Read OpenRASE: Service Function Chain Emulation by Theviyanthan Krishnamohan, Paul Harvey (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-22131.txt.

Introduction

Theviyanthan Krishnamohan and Paul Harvey release a playground for chained network services. Their 2025 paper presents OpenRASE, an emulator for service function chains, the linked virtual tools that steer modern traffic. Testing resource algorithms for such chains has been hampered by fake, stiff, or tiny sandboxes. Their build marries Mininet networking with Docker containers.

That pairing lets real CPU use and delay be measured as conditions shift. Two allocation algorithms, one of them a live genetic search, were tried inside to prove the rig. The authors argue that honest emulation is the missing step between clever chain algorithms and networks that run themselves.

Catalog record

CREATORTheviyanthan Krishnamohan, Paul Harvey
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE26 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2507-22131

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

OpenRASE: Service Function Chain Emulation Theviyanthan Krishnamohan, Paul Harvey Abstract Service Function Chains (SFCs) are one of the key enablers in providing programmable computer networks, paving the way for network autonomy. However, this also introduces new challenges, such as resource allocation and optimisation related to their operation, requiring new algorithms to address these challenges. Various tools have been used in the literature to evaluate these algorithms. However, these tools suffer from inaccuracy, low fidelity, unscalability, inflexibility, or additional code requirements. This paper introduces an emulator based on Mininet and Docker for SFCs called OpenRASE . The goal of OpenRASE is to enable the exploration of resource allocation algorithms for SFCs in a dynamic setting, allowing real CPU usage and latency to be measured. We describe the design and implementation of OpenRASE and discuss its characteristics. We also experimentally evaluate two different algorithms to address the SFC resource allocation challenge, including an online Genetic Algorithm, using OpenRASE to show its effectiveness and practicality for dynamic network conditions. \DeclareSourcemap \maps \map \step [ fieldsource=doi, match= \regexp { _}—{_}— _, replace= \regexp _ ] \map \step [ fieldsource=url, match= \regexp { _}—{_}— _, replace= \regexp _ ] I Introduction Computer networks are an essential part of everyday life, playing a major role in our economy, education, safety, and especially during COVID-19, our mental health [ 1 ] . As a result, it is essential to ensure their continued high-quality operation. Enabling our networks to operate more autonomously is one path to support this, which can lead to cost savings, quality of service increases, more energy-efficient operation, and scaling beyond current human capacity [ 2 ] . One of the key technological enablers on this path is network virtualisation via Service Function Chains (SFCs). SFCs deploy network functions, such as firewalls, virtually on servers, and establish network links between them using Software Defined Networking (SDN). In effect, SFCs introduce a virtual service overlay over a physical network. Despite their flexibility, SFCs also introduce new challenges [ 3 ] , such as how to compose the structure of an SFC (Chain Composition problem), how to allocate network, computational, and storage resources to deploy an SFC’s network functions (Forwarding Graph Embedding problem), and how to schedule the SFC’s network functions (Scheduling problem). These challenges are collectively termed the Network Function Virtualisation Resource Allocation (NFV-RA) problem [ 3 ] . From the literature, there are several approaches to generating solutions to the NFV-RA problem (referred to as solver in the remainder of the paper), as well as several different tools and techniques to evaluate these solutions (Section II ). Unfortunately, currently available evaluation tools either lack sufficient fidelity, especially for dynamic network environments, or do not support NFV-RA evaluation natively. This limits their effectiveness in providing a robust evaluation of NFV-RA solvers. This is particularly challenging when trying to address the NFV-RA problem for real network environments that exhibit runtime dynamic behaviour (Section VII ). To address this challenge, we introduce Open R esource A llocation to S FC E mulator (OpenRASE). OpenRASE (Section III ) is an emulation environment for SFCs based on the Mininet network emulator and Docker virtualisation. To the best of the authors’ knowledge, OpenRASE is the first emulator specifically designed to support emulated NFV-RA experiments. To characterise the capabilities of OpenRASE, we compare it against ALEVIN, the most popular tool in the literature for developing solutions to the NFV-RA problem, by experimentally using an ALEVIN solver in OpenRASE. We also demonstrate experimental evaluation of a Genetic Algorithm-based solver to show the versatility of OpenRASE. II Related Work A review of the literature related to the NFV-RA problem domain between 2014 and 2025 was performed to examine the tools used to evaluate the NFV-RA algorithms. The tools found in the literature can be categorised as simulators, emulators, testbeds, and production-grade virtualisation tools. Simulators are software programs that model reality at a reduced level of accuracy. They provide a faster, scalable, flexible and cost-effective way of testing solutions compared to emulators and testbeds. However, they lack fidelity and, as a result, may not provide an accurate evaluation [ 4 ] . Testbeds provide a controlled physical environment based on the operational or deployment environment and can provide a very high-fidelity evaluation. However, they are expensive, require maintenance, and are less flexible and scalable compared to simulation [ 4 ] . Emulators exist in the continuum between simulators and testbeds [ 4 ] . Instead of modelling or abstracting reality, they duplicate reality, providing higher fidelity than simulators and higher flexibility and scalability than testbeds. Like testbeds, emulators can express uncertainty in results, leading to more realistic and variable behaviour during evaluation cycles. This often comes at the cost of higher execution times compared to simulators. Production-grade virtualisation tools (PGVTs) , such as OpenStack and Kubernetes, have also been used to evaluate solutions to the NFV-RA problem. Given these are virtualisation tools, it can be argued that they are similar to emulators in that they are more flexible and scalable than testbeds, while offering higher fidelity than simulators. However, they do not support NFV-RA scenarios natively, thus requiring additional code. A summary of the tools in the literature can be found in Table I . It can be seen that most works use simulators in developing NFV-RA algorithms. The use of a simulator like ALEVIN, the most popular tool in the literature, abstracts away challenges associated with dynamic runtime environments, also known as online experimentation (Section VII ), such as time consumption and an uncertain environment. As a result, solutions to the NFV-RA problem may not operate effectively after deployment in production systems. TABLE I: Summary of the simulators, emulators, testbeds, and Production-Grade Virtualisation Tools found in the literature. Name Type Supports NFV-RA Frequency ALEVIN Simulator Forwarding Graph Embedding Only 12 Mininet Emulator No 7 CloudSim Simulator No 2 OpenStack PGVT No 5 Kubernetes PGVT No 2 OpenNetVM PGVT No 2 OMNet++ Simulator No 2 CREATE-NET Testbed No 1 5GinFIRE Testbed No 1 RPi platform Testbed No 1 At the same time, as experiments often require frequent changes to the underlying substrate network, testbeds present challenges of scalability and flexibility. Emulators and PGVTs are flexible and scalable and do not abstract away the challenges associated with dynamic runtime environments. However, they require additional code to support evaluating NFV-RA algorithms. III OpenRASE To enable accurate online NFV-RA experiments, we now introduce OpenRASE. OpenRASE is an emulator developed to evaluate NFV-RA solvers and is available online 1 1 1 https://github.com/Project-Kelvin/OpenRASE . III-A Design Principles III-A 1 Standards Compliant OpenRASE is based on the European Telecommunication Standard Institute (ETSI)’s [ 5 ] NFV architecture and Internet Engineering Task Force (IETF)’s [ 6 ] SFC architecture. The NFV architecture provides a framework for deploying and managing Virtual Network Functions (VNFs), while the SFC architecture is used to create and maintain SFCs on networks. III-A 2 Python, Mininet, and Docker-based OpenRASE is built on top of Containernet [ 7 ] , a fork of Mininet that enables Docker containers to be used as Mininet hosts. It uses these Docker containers to emulate servers (also referred to as hosts in this paper) and Mininet to establish network connections among Docker containers.

Read the complete text