# Policy-Based Composition and Embedding of Extended Virtual Networks and SFCs for IIoT

^{*}

^{†}

## Abstract

**:**

## 1. Introduction

## 2. Related Work

## 3. Problem Description

#### 3.1. Use Case

#### 3.2. System Model

- A physical topology $SN({N}_{SN},{E}_{SN})$ is defined by the service provider.
- The application requirements $AR({N}_{AR},{E}_{AR})$ are defined by the user.
- A set of policies (defined by the service provider), where each policy is composed of:
- –
- A pattern P: conditions on the $SN$ and $EVN$ for checking and comparison of properties, demands, and resources.
- –
- Transformation T: if P match is found, then perform $T(EV{N}_{i-1},SN)=EV{N}_{i}$, where $EV{N}_{0}=AR$, and T is a set of rules.
- –
- A rule R adds or copies virtual links or nodes, adds or changes the properties or demands, or adds VNFs to the VNF demand of a VLi.

- The policies are applied with the predefined order on the input AR and SN.
- EVN mapping:
- –
- For each VLi in the final EVN, if the VLi has a VNF demand, perform topological sorting on the VNF demand $TS(vnfDemand,vnfDependencies)=List\phantom{\rule{0.277778em}{0ex}}of\phantom{\rule{0.277778em}{0ex}}FG$, where the output is a set of FGs.
- –
- Map EVN nodes on the SN based on their properties and demands.
- –
- Map EVN simple VLis based on their properties and demands.
- –
- For each VLi with a VNF demand, map a selected FG while using the chain embedding proactive heuristic.

Algorithm 1: System Model |

#### 3.3. General Dependency Graph

## 4. Methodology

#### 4.1. Transformations

#### 4.1.1. Redundancy

#### 4.1.2. Nodes Types

#### 4.1.3. Security VNFs

#### 4.1.4. Low Latency VNF

#### 4.1.5. Bandwidth

#### 4.1.6. CPU and Load Balancing

#### 4.2. Creating FGs Using Topological Sorting

Algorithm 2: Topological sorting |

Algorithm 3: Finding all topological sortings |

#### 4.3. EVN Embedding

#### 4.4. Chain Embedding Heuristic

Algorithm 4: Proactive chain embedding algorithm |

#### 4.5. Complexity of the Proposed Methods

- Neglecting the operations applied by the rule on only a small subset of the EVN entities.
- Considering that the maximum size of the final EVN (nodes M, links L) is known and used to estimate the complexity even for intermediate EVNs.
- (D, D) is the size of the dependency graph. We consider that, in NFV, a function typically has one dependency, so the number of edges in the dependency graph can be approximated to the number of VNFs. The size of the dependency graph is considered the size of each VNF demand.
- We consider that finding one topological sorting of the VNF demand (one FG) is enough, since we do not yet classify the resulting FGs or trying to select the best FG based on a specific criteria.
- P is the diameter of the SN graph (the number of edges in the shortest path between the most distant vertices). We use this worst-case value as the length of the shortest path on which the chain will be embedded.
- We are calculating the complexities in the worst cases without considering locations and domains that might reduce the search space for node and link mapping, in particular, in specific topologies, like in our use case.

- Transformations:
- –
- Redundancy—O(L): checking the redundancy demand of all EVN links and copying them when needed.
- –
- Node types—O(N.M): comparing certain properties of SNos and VNos and adding the required entities to the EVN with their demands.
- –
- Security—O(L): comparing the source and destination locations/domains of the EVN links and adding the required VNFs.
- –
- Low latency—O(L): checking the latency demand of the EVN links and adding the TAS VNF where needed.
- –
- Bandwidth—O(L): checking the bandwidth demand of each VLi and adjusting it based on the bandwidth demands of the incoming edges to its source VNo.
- –
- CPU—O(M): adjusting the CPU demand of all EVN nodes based on the incoming bandwidth, cloning certain nodes, and then adding load balancing function.
- –
- Total complexity of the transformations: O(4L+M (1+N)) = O(L + M.N) when considering that $N\gg 1$.

- Topological sorting—O(L.D) = O(L): the complexity of finding one topological sorting is O(D+D) = O(D). We assume here that D is small and fixed and that we need to calculate a FG for each VLi.
- Node mapping—O (N.M): finding the SNo with the matching ID, properties, or resources of each VNo.
- Chain embedding—O (2P)=O(P): when mapping a VNF is tried on a SNo, the demands are compared to the resources. A VNF can be mapped on the latest SNo used from the path or any other following node. In the worst case, all path SNos are checked twice for each placed VNF and next VNF that cannot be placed on the same SNo, so a following one is checked. The P value varies with the SN size.
- Link mapping: the complexity of the k-shortest path algorithm is O(E+N.log N)) [21]. This algorithm is used for mapping all VLis. We assume for the worst case that all VLis have VNF demands for which chain embedding will be performed. The total complexity of link mapping is O(L.(E+N.log N+P)).

## 5. Results

#### 5.1. Implementation

#### 5.2. Applying the Methods to the Use Case

#### 5.3. Evaluation of the Chain Embedding Using a Random Topology

- Initial number of nodes: ${m}_{0}$ = 1
- Number of time steps: 20
- Number of new edges per time step: 1

#### 5.3.1. Runtime

#### Runtime for Increasing SN Size

#### Runtime for Increased Number of EVNs

#### 5.3.2. Comparison to LightChain

#### Acceptance Ratio with Increasing FG Length

#### Average Path Utilization with Increasing FG Length

#### Acceptance Ratio for Increasing CPU Capacity

## 6. Discussion

## Author Contributions

## Funding

## Conflicts of Interest

## References

- Fischer, A.; Botero, J.F.; Beck, M.T.; de Meer, H.; Hesselbach, X. Virtual Network Embedding: A Survey. IEEE Commun. Surv. Tutor.
**2013**, 15, 1888–1906. [Google Scholar] [CrossRef] - Herrera, J.G.; Botero, J.F. Resource Allocation in NFV: A Comprehensive Survey. IEEE Trans. Netw. Serv. Manag.
**2016**, 13, 518–532. [Google Scholar] [CrossRef] - Mehraghdam, S.; Keller, M.; Karl, H. Specifying and placing chains of virtual network functions. In Proceedings of the 2014 IEEE 3rd International Conference on Cloud Networking (CloudNet), Luxembourg, 8–10 October 2014; pp. 7–13. [Google Scholar] [CrossRef][Green Version]
- Ocampo, A.F.; Gil-Herrera, J.; Isolani, P.H.; Neves, M.C.; Botero, J.F.; Latré, S.; Zambenedetti, L.; Barcellos, M.P.; Gaspary, L.P. Optimal Service Function Chain Composition in Network Functions Virtualization. In Security of Networks and Services in an All-Connected World, Proceedings of the 11th IFIP WG 6.6 International Conference on Autonomous Infrastructure, Management, and Security, AIMS 2017, Zurich, Switzerland, 10–13 July 2017; Springer International Publishing: Cham, Switzerland, 2017; pp. 62–76. [Google Scholar] [CrossRef][Green Version]
- Beck, M.T.; Botero, J.F. Coordinated Allocation of Service Function Chains. In Proceedings of the 2015 IEEE Global Communications Conference (GLOBECOM), San Diego, CA, USA, 6–10 December 2015; pp. 1–6. [Google Scholar] [CrossRef]
- Hirwe, A.; Kataoka, K. LightChain: A lightweight optimisation of VNF placement for service chaining in NFV. In Proceedings of the 2016 IEEE NetSoft Conference and Workshops (NetSoft), Seoul, Korea, 6–10 June 2016; pp. 33–37. [Google Scholar] [CrossRef]
- Wang, Z.; Zhang, J.; Huang, T.; Liu, Y. Service Function Chain Composition, Placement, and Assignment in Data Centers. IEEE Trans. Netw. Serv. Manag.
**2019**, 16, 1638–1650. [Google Scholar] [CrossRef] - Li, D.; Hong, P.; Xue, K.; Pei, J. Virtual Network Function Placement Considering Resource Optimization and SFC Requests in Cloud Datacenter. IEEE Trans. Parallel Distrib. Syst.
**2018**, 1. [Google Scholar] [CrossRef] - Wang, Y.; Li, Z.; Xie, G.; Salamatian, K. Enabling automatic composition and verification of service function chain. In Proceedings of the 2017 IEEE/ACM 25th International Symposium on Quality of Service (IWQoS), Vilanova i la Geltru, Spain, 14–16 June 2017; pp. 1–5. [Google Scholar]
- Li, H.; Wang, L.; Wen, X.; Lu, Z.; Ma, L. Constructing Service Function Chain Test Database: An Optimal Modeling Approach for Coordinated Resource Allocation. IEEE Access
**2018**, 6, 17595–17605. [Google Scholar] [CrossRef] - Wang, M.; Cheng, B.; Zhao, S.; Li, B.; Feng, W.; Chen, J. Availability-Aware Service Chain Composition and Mapping in NFV-Enabled Networks. In Proceedings of the 2019 IEEE International Conference on Web Services (ICWS), Milan, Italy, 8–13 July 2019; pp. 107–115. [Google Scholar]
- Beck, M.T.; Linnhoff-Popien, C.; Fischer, A.; Kokot, F.; de Meer, H. A simulation framework for Virtual Network Embedding algorithms. In Proceedings of the 2014 16th International Telecommunications Network Strategy and Planning Symposium (Networks), Funchal, Portugal, 17–19 September 2014; pp. 1–6. [Google Scholar] [CrossRef]
- Beck, M.T.; Botero, J.F.; Samelin, K. Resilient allocation of service Function chains. In Proceedings of the 2016 IEEE Conference on Network Function Virtualization and Software Defined Networks (NFV-SDN), Palo Alto, CA, USA, 7–10 November 2016; pp. 128–133. [Google Scholar] [CrossRef]
- 5 Top Industrial IoT Use Cases. 2017. Available online: https://www.ibm.com/blogs/internet-of-things/top-5-industrial-iot-use-cases/ (accessed on 14 September 2020).
- Aazam, M.; Zeadally, S.; Harras, K.A. Deploying fog computing in industrial internet of things and industry 4.0. IEEE Trans. Ind. Inf.
**2018**, 14, 4674–4682. [Google Scholar] [CrossRef] - Rozenberg, G. Handbook of Graph Grammars and Computing by Graph Transformation; World Scientific: Singapore, 1997. [Google Scholar]
- Fischer, A.; Kühn, R.; Mandarawi, W.; de Meer, H. Modeling Security Requirements for VNE Algorithms. In Proceedings of the 10th EAI International Conference on Performance Evaluation Methodologies and Tools, Taormina, Italy, 25–28 October 2016; pp. 149–154. [Google Scholar] [CrossRef][Green Version]
- Mandarawi, W.; Chahed, H.; de Meer, H. A Framework for Virtualizing Time-aware Shaper Using High Performance NFV. In Proceedings of the IEEE 25th International Conference on Emerging Technologies and Factory Automation, ETFA 2020, Vienna, Austria, 8–11 September 2020. [Google Scholar]
- Chen, X.; Zhang, D.; Wang, X.; Zhu, K.; Zhou, H. P4SC: Towards High-Performance Service Function Chain Implementation on the P4-Capable Device. In Proceedings of the 2019 IFIP/IEEE Symposium on Integrated Network and Service Management (IM), Arlington, VA, USA, 8–12 April 2019; pp. 1–9. [Google Scholar]
- Ottmann, T.; Widmayer, P. Algorithmen und Datenstrukturen, 4. Auflage; Spektrum Akademischer Verlag: Berlin/Heidelberg, Germany, 2002. [Google Scholar]
- Eppstein, D. Finding the k shortest paths. SIAM J. Comput.
**1998**, 28, 652–673. [Google Scholar] [CrossRef][Green Version] - Waxman, B.M. Routing of multipoint connections. IEEE J. Sel. Areas Commun.
**1988**, 6, 1617–1622. [Google Scholar] [CrossRef] - Albert, R.; Barabási, A.L. Statistical mechanics of complex networks. Rev. Mod. Phys.
**2002**, 74, 47–97. [Google Scholar] [CrossRef][Green Version]

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## Share and Cite

**MDPI and ACS Style**

Mandarawi, W.; Rottmeier, J.; Rezaeighale, M.; de Meer, H. Policy-Based Composition and Embedding of Extended Virtual Networks and SFCs for IIoT. *Algorithms* **2020**, *13*, 240.
https://doi.org/10.3390/a13090240

**AMA Style**

Mandarawi W, Rottmeier J, Rezaeighale M, de Meer H. Policy-Based Composition and Embedding of Extended Virtual Networks and SFCs for IIoT. *Algorithms*. 2020; 13(9):240.
https://doi.org/10.3390/a13090240

**Chicago/Turabian Style**

Mandarawi, Waseem, Jürgen Rottmeier, Milad Rezaeighale, and Hermann de Meer. 2020. "Policy-Based Composition and Embedding of Extended Virtual Networks and SFCs for IIoT" *Algorithms* 13, no. 9: 240.
https://doi.org/10.3390/a13090240