Elastic Manufacturing Process Landscapes

Abstract

Because of increasing competition and cost pressure, the manufacturing industry is currently undergoing massive changes that are facilitated by the usage of Information Technologies. Two particular aspects are the usage of Business Process Management (BPM) and Cloud technologies concepts in the manufacturing domain. Rapid elasticity is crucial for the enactment of manufacturing processes in the Cloud. This work in progress paper aims at presenting some basic principles of elastic processes in the manufacturing domain. Henceforth, an approach towards adaptive infrastructure provisioning that allows for predefined Quality of Service (QoS) and Service Level Agreement (SLA) metrics in manufacturing Cloud environments is considered.

Keywords

• Olena Skarlat
• Elastic Processes
• Cloud Manufacturing
• Cyber-Physical System
• Industry 4
• 0
• Cloud Computing

-

Introduction s e c i v r e S

Manufacturing Process N

Manufacturing Process A t1 Status: Ok Delivery: Ok t2 Status: Late Delivery: At risk t3 Status: Prepared (waiting)

Manufacturing Company

t4 Status: Prepared (waiting)

Manufacturing

Company:

Headquarters

The Cloud Manufacturing Platform

Supplier A

Supplier B

Manufacturing

Company:

Plant

is a motivational scenario of how to apply elasticity principles in manufacturing process landscapes and an overview of instrumentation for its implementation, specifically, of service scheduling and resource provisioning mechanisms. 2

Scenario and Research Questions To motivate our work we consider a scenario from the manufacturing industry called Cloud Manufacturing. Cloud Manufacturing is a new concept of networked manufacturing that makes use of crowdsourcing and outsourcing models for manufacturing processes. Manufacturing processes here has to be considered as a set of process steps to be performed to create a certain manufacturing product. These process steps are in reality single services, which are responsible for a certain manufacturing asset on the shop floor. Ideally, Cloud Manufacturing offers means to integrate single services of the manufacturing processes from distributed locations as if the complete manufacturing was carried out on the same shop floor (Fig. 1). For this the manufacturers virtualize their single services of manufacturing processes. An integration is possible via a Cloud Manufacturing platform, where these services are presented, advertized, leased, and sold as a part of manufacturing processes maintained in the platform.

The existence and popularization of virtual enterprises sets the challenges to this new concept of Cloud Manufacturing. Virtual enterprises imply plugging together independent virtual factories to manufacture a certain product [ 7,2 ]. In contrast, Cloud Manufacturing assumes encapsulation of manufacturing assets into services in the Cloud Manufacturing platform as an inevitable part of its scenario. Cloud Manufacturing provides means to abstract single manufacturing assets, like sensors and Cyber-Physical System objects, and present them as services in the Cloud Manufacturing platform. The aim is to inform about available services on the Marketplace, suggest appropriate services or needed substitutes to the manufacturers, and optimize manufacturing processes.

Manufacturer inMvoakneusfacturern

Cloud Manufacturing Platform manages schedules invokes

Manufacturing Process Manufacturing Process n

Service

Service

Service eBPMS

VM

Assignment Text

Assignment Text Assignment n Text allocates

VM n

VM Computational Resources The challenge here is that thousands of manufacturers must be simultaneously served. Correspondingly, a large amount of interdependent processes with different QoS and SLA demands may be requested at any point of time. Therefore, elastic processes are a promising approach in Cloud Manufacturing [ 5 ]. To adhere to these principles, resource elasticity allowing on-demand scaling of computational resources has to be established inside the Cloud Manufacturing environment. However, computational resource elasticity is not the only dimension to be regarded: cost and quality elasticity bring flexibility in price levels of Cloud services and close the tradeoff between QoS metrics and cost. To enact elastic processes, a BPMS with features to control the Cloud is needed, as depicted in (Fig. 2). Allowing for the demand in computational resources and taking into account QoS demands, such an eBPMS schedules process instances (respectively, the single services used for the enactment of these processes) and allocates Cloudbased resources as necessary. To correlate with elasticity metrics (resources, quality, and cost), it is assumed that the manufacturing processes mentioned in the scenario are composed from single software services instantiated on Virtual Machines (VMs) in the Cloud. With this in mind, achieving elasticity implies leasing VMs when needed, deployment of services onto those VMs, invocation of service instances using a calculated schedule, and releasing resources after services are finished.

In real-world manufacturing processes an eBPMS needs to be able to solve these optimization problems in very short time and under potentially heavy load. Existing exact methods can provide a solution for small-scale scenarios. However, applying exact methods in large-scale manufacturing process landscapes is time-consuming or may not provide any solution in polynomial time, since the underlying decision problem is NP-hard [ 3 ]. Therefore an elastic manufacturing process enactment requires reasoning methodologies based on heuristic algorithms. The research question that is tackled in this work is the following: “What is an appropriate reasoning model for smart resource provisioning in elastic manufacturing process landscapes, and what are the methodologies and instrumentation to support an application of this model in the manufacturing domain?” Specifically, this work at progress is focused on presenting an underlying optimization model, and on implementing a heuristic reasoning mechanism for an eBPMS to adaptively select services for service orchestration in manufacturing processes, optimizing service scheduling and resource allocation.

To conclude, the use of elastic processes within the Cloud Manufacturing domain facilitates machine- and human-collaboration in the manufacturing industry. The aim of Cloud Manufacturing is to perform a transformation from production-oriented manufacturing processes to service-oriented manufacturing process networks by virtualizing manufacturing assets as services similarly as Software-as-a-Service or Platform-as-a-Service solutions are already provided by the Cloud providers. Clouds bring benefits to the manufacturing process management, and adaptive elastic enactment of manufacturing processes is intended to take into account QoS and SLA metrics and to perform an optimization and a runtime adjustment of infrastructural components.

Acknowledgments This paper is supported by TU Wien research funds. This work is partially supported by the Commission of the European Union within the CREMA H2020-RIA project (Grant agreement no. 637066).

References

1. Dustdar , S. , Guo , Y. , Satzger , B. , Truong , H.L. : Principles of elastic processes . IEEE Internet Computing 15 ( 2 ), 66 - 71 ( 2011 )

2. Hao , Y. , Shamsuzzo , A. , Helo , P. : Designing of virtual factory information system by enterprise portal . In: Proc. 17th IEEE International Enterprise Distributed Object Computing Conference Workshops . pp. 258 - 266 . IEEE (9-13 Sep 2013 )

3. Hoenisch , P. , Schuller , D. , Schulte , S. , Hochreiner , C. , Dustdar , S. : Optimization of Complex Elastic Processes (accepted for publication) . IEEE Transactions on Services Computing ( 2015 )

4. Leitner , P. , Hummer , W. , Dustdar , S. : Cost-based optimization of service compositions . IEEE Transactions on Services Computing 6 ( 2 ), 239 - 251 ( 2013 )

5. Schulte , S. , Hoenisch , P. , Hochreiner , C. , Dustdar , S. , Klusch , M. , Schuller , D. : Towards process support for cloud manufacturing . In: Proc. 18th IEEE International Enterprise Distributed Object Computing Conference . pp. 142 - 149 . IEEE Computer Society (1-4 Sep 2014 )

6. Schulte , S. , Janiesch , C. , Venugopal , S. , Weber , I. , Hoenisch , P. : Elastic Business Process Management: State of the Art and Open Challenges for BPM in the Cloud . Future Generation Computer Systems 46 , 36 - 50 ( 2015 )

7. Schulte , S. , Schuller , D. , Steinmetz , R. , Abels , S. : Plug-and-play virtual factories . IEEE Internet Computing 16 , 78 - 82 ( 2012 )

8. Wen , X. , Zhou , X. : Servitization of manufacturing industries based on cloud-based business model and the down-to-earth implementary path . International Journal of Advanced Manufacturing Technology pp. 1 - 18 ( 2014 )

9. Wu , D. , Greer , M.J. , Rosen , D.W. , Schaefer , D. : Cloud manufacturing: drivers, current status, and future trends . In: Proc. ASME International Manufacturing Science and Engineering Conference . pp. 1 - 10 . ASME ( 10 - 14 Jun 2013 )