Pub Date : 2019-11-01DOI: 10.1109/CloudNet47604.2019.9064140
Prasad Talasila, D. Lucani
This paper proposes generalized deduplication, a concept where similar data is systematically deduplicated by first transforming chunks of each file into two parts: a basis and a deviation. This increases the potential for compression as more chunks can have a common basis that can be deduplicated by the system. The deviation is kept small and stored together with an identifier to its chunk, e.g., hash of a chunk, in order to recover the original data without errors or distortions. This paper characterizes the performance of generalized deduplication using Golomb-Rice codes as a suitable data transform function to discover similarities across all files stored in the system. Considering different synthetic data distributions, we show in theory and simulations that generalized deduplication can result in compression factors of 300 (high compression), i.e., 300 times less storage space, and that this compression is achieved with 60,000 times fewer data chunks inserted into the system compared to classic deduplication (compression gains start earlier). Finally, we show that the table/registry to recognize similar chunks is 10,000 times smaller for generalized deduplication compared to the table in classic deduplication techniques, which will result in less RAM usage in the storage system.
{"title":"Generalized Deduplication: Lossless Compression by Clustering Similar Data","authors":"Prasad Talasila, D. Lucani","doi":"10.1109/CloudNet47604.2019.9064140","DOIUrl":"https://doi.org/10.1109/CloudNet47604.2019.9064140","url":null,"abstract":"This paper proposes generalized deduplication, a concept where similar data is systematically deduplicated by first transforming chunks of each file into two parts: a basis and a deviation. This increases the potential for compression as more chunks can have a common basis that can be deduplicated by the system. The deviation is kept small and stored together with an identifier to its chunk, e.g., hash of a chunk, in order to recover the original data without errors or distortions. This paper characterizes the performance of generalized deduplication using Golomb-Rice codes as a suitable data transform function to discover similarities across all files stored in the system. Considering different synthetic data distributions, we show in theory and simulations that generalized deduplication can result in compression factors of 300 (high compression), i.e., 300 times less storage space, and that this compression is achieved with 60,000 times fewer data chunks inserted into the system compared to classic deduplication (compression gains start earlier). Finally, we show that the table/registry to recognize similar chunks is 10,000 times smaller for generalized deduplication compared to the table in classic deduplication techniques, which will result in less RAM usage in the storage system.","PeriodicalId":340890,"journal":{"name":"2019 IEEE 8th International Conference on Cloud Networking (CloudNet)","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129216051","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-11-01DOI: 10.1109/CloudNet47604.2019.9064128
Takaaki Sawa, Fujun He, Akio Kawabata, E. Oki
This paper proposes an algorithm for the distributed server allocation problem, namely Minimizing the Maximum Delay (MMD), where an optimal solution is obtained when all server-server delays are the same constant value. We prove that MMD obtains an optimal solution with the polynomial time complexity.
{"title":"Polynomial-time Algorithm for Distributed Server Allocation Problem","authors":"Takaaki Sawa, Fujun He, Akio Kawabata, E. Oki","doi":"10.1109/CloudNet47604.2019.9064128","DOIUrl":"https://doi.org/10.1109/CloudNet47604.2019.9064128","url":null,"abstract":"This paper proposes an algorithm for the distributed server allocation problem, namely Minimizing the Maximum Delay (MMD), where an optimal solution is obtained when all server-server delays are the same constant value. We prove that MMD obtains an optimal solution with the polynomial time complexity.","PeriodicalId":340890,"journal":{"name":"2019 IEEE 8th International Conference on Cloud Networking (CloudNet)","volume":"38 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121361542","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-11-01DOI: 10.1109/CloudNet47604.2019.9064145
Rasoul Behravesh, Estefanía Coronado, D. Harutyunyan, R. Riggio
With the advent of 5G systems, telecommunication service providers (TSPs) have been facing a tremendous transition by the raised expectations of supporting billions of IoT devices and an unprecedented amount of generated data. This revolutionary transformation necessitates innovative approaches such as multi-access edge computing (MEC) to meet the requirements of many novel applications in terms of their high data rate and low latency. The idea behind MEC is to move data, virtualization, and processing capabilities from central data centers to the edge of the network. However, resources at the network edge are very scarce and costly to provision. Therefore, TSPs have to make smart decisions on how to utilize the network resources such as to make sure that the user service requirements (e.g., data rate, latency) are satisfied while the network resources are used most efficiently. In this paper, we study the problem of joint user association, VNF placement, and resource allocation, employing mixed-integer linear programming (MILP) technique. The objectives of the formulations are to minimize (i) the service provisioning cost, (ii) the number of VNF instances, and (iii) the transport network utilization, having an overarching goal of drawing a comparison between these different approaches.
{"title":"Joint User Association and VNF Placement for Latency Sensitive Applications in 5G Networks","authors":"Rasoul Behravesh, Estefanía Coronado, D. Harutyunyan, R. Riggio","doi":"10.1109/CloudNet47604.2019.9064145","DOIUrl":"https://doi.org/10.1109/CloudNet47604.2019.9064145","url":null,"abstract":"With the advent of 5G systems, telecommunication service providers (TSPs) have been facing a tremendous transition by the raised expectations of supporting billions of IoT devices and an unprecedented amount of generated data. This revolutionary transformation necessitates innovative approaches such as multi-access edge computing (MEC) to meet the requirements of many novel applications in terms of their high data rate and low latency. The idea behind MEC is to move data, virtualization, and processing capabilities from central data centers to the edge of the network. However, resources at the network edge are very scarce and costly to provision. Therefore, TSPs have to make smart decisions on how to utilize the network resources such as to make sure that the user service requirements (e.g., data rate, latency) are satisfied while the network resources are used most efficiently. In this paper, we study the problem of joint user association, VNF placement, and resource allocation, employing mixed-integer linear programming (MILP) technique. The objectives of the formulations are to minimize (i) the service provisioning cost, (ii) the number of VNF instances, and (iii) the transport network utilization, having an overarching goal of drawing a comparison between these different approaches.","PeriodicalId":340890,"journal":{"name":"2019 IEEE 8th International Conference on Cloud Networking (CloudNet)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129143706","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-11-01DOI: 10.1109/CloudNet47604.2019.9064111
Yoonhyeong Lee, Hyunseok Choi, Youngju Nam, Sungjin Park, Euisin Lee
Recently, Vehicular Ad hoc Networks (VANETs) introduce a Vehicular Cloud (VC) model for collaborating to share and use resources of vehicles to create value-added services. To construct a VC, a vehicle should search vehicles that intend to provide their own resource. The single-hop search cannot search enough provider vehicles due to a small coverage and out-of-sight-zone of communications. On the other hand, the multi-hop search causes very high traffics for large coverage searching and frequent connection breakages. Recently, many Roadside Units (RSUs) have been deployed on roads to collect the information of vehicles in their own coverages and connect them to Internet. Thus, we propose a RSU-aided vehicular resource search and cloud construction mechanism in VANETS. In the proposed mechanism, a RSU selects provider vehicles enabled to provide resources needed for constructing a VC based on the information of location and mobility of vehicles. In the proposed mechanism, the criteria for determining provider vehicles are the connection duration between each candidate vehicle and the requester vehicle, the resource size of each candidate vehicle, and its connection starting time to the requester vehicle. Simulation results verify that the proposed mechanism achieves better performance than the existing mechanism.
{"title":"RSU-driven Cloud Construction and Management Mechanism in VANETs","authors":"Yoonhyeong Lee, Hyunseok Choi, Youngju Nam, Sungjin Park, Euisin Lee","doi":"10.1109/CloudNet47604.2019.9064111","DOIUrl":"https://doi.org/10.1109/CloudNet47604.2019.9064111","url":null,"abstract":"Recently, Vehicular Ad hoc Networks (VANETs) introduce a Vehicular Cloud (VC) model for collaborating to share and use resources of vehicles to create value-added services. To construct a VC, a vehicle should search vehicles that intend to provide their own resource. The single-hop search cannot search enough provider vehicles due to a small coverage and out-of-sight-zone of communications. On the other hand, the multi-hop search causes very high traffics for large coverage searching and frequent connection breakages. Recently, many Roadside Units (RSUs) have been deployed on roads to collect the information of vehicles in their own coverages and connect them to Internet. Thus, we propose a RSU-aided vehicular resource search and cloud construction mechanism in VANETS. In the proposed mechanism, a RSU selects provider vehicles enabled to provide resources needed for constructing a VC based on the information of location and mobility of vehicles. In the proposed mechanism, the criteria for determining provider vehicles are the connection duration between each candidate vehicle and the requester vehicle, the resource size of each candidate vehicle, and its connection starting time to the requester vehicle. Simulation results verify that the proposed mechanism achieves better performance than the existing mechanism.","PeriodicalId":340890,"journal":{"name":"2019 IEEE 8th International Conference on Cloud Networking (CloudNet)","volume":"45 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121558772","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-11-01DOI: 10.1109/CloudNet47604.2019.9064117
R. Kang, Fujun He, Takehiro Sato, E. Oki
This paper proposes an optimization model to derive a virtual network function (VNF) allocation for time slots in sequence aiming to maximize the continuous available time of service function chains (SFCs) in a network. The proposed model suppresses service interruptions due to the unavailability of nodes and the reallocation of VNFs. Compared with conventional models, the proposed model computes the VNF allocation in a series of time slots based on the network node availability schedule, which provides information on whether each network node is available at each time slot. We formulate the proposed model as an integer linear programming problem with the goal of maximizing the minimum number of longest continuous available time slots in each SFC. The numerical results show that the proposed model improves the continuous available time of SFCs compared with existing models.
{"title":"Virtual Network Function Allocation to Maximize Continuous Available Time of Service Function Chains","authors":"R. Kang, Fujun He, Takehiro Sato, E. Oki","doi":"10.1109/CloudNet47604.2019.9064117","DOIUrl":"https://doi.org/10.1109/CloudNet47604.2019.9064117","url":null,"abstract":"This paper proposes an optimization model to derive a virtual network function (VNF) allocation for time slots in sequence aiming to maximize the continuous available time of service function chains (SFCs) in a network. The proposed model suppresses service interruptions due to the unavailability of nodes and the reallocation of VNFs. Compared with conventional models, the proposed model computes the VNF allocation in a series of time slots based on the network node availability schedule, which provides information on whether each network node is available at each time slot. We formulate the proposed model as an integer linear programming problem with the goal of maximizing the minimum number of longest continuous available time slots in each SFC. The numerical results show that the proposed model improves the continuous available time of SFCs compared with existing models.","PeriodicalId":340890,"journal":{"name":"2019 IEEE 8th International Conference on Cloud Networking (CloudNet)","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128539775","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-11-01DOI: 10.1109/CloudNet47604.2019.9064108
C. Morin, Géraldine Texier, C. Caillouet, Gilles Desmangles, Cao-Thanh Phan
The Network Functions Virtualisation (NFV) concept offers network operators the ability to provide more scalable and less expensive services, free from the limitations inherent to hardware devices. However, in 5G networks, the functions must be deployed not only in large central data centers, but also in the edge. We propose an algorithm that solves the Virtual Network Function Chain Placement Problem allowing a fine management of these rare resources in order to respond to the greatest number of requests possible. Because networks can be divided into several entities belonging to different tenants who are reluctant to reveal their internal topologies, we propose a heuristic that allows the NFV orchestrator to place the function chains based only on an abstract view of the infrastructure network. We leverage this approach to address the complexity of the problem in large mono- or multi-tenant networks. We analyze the efficiency of our algorithm and heuristic with respect to a wide range of parameters and topologies.
{"title":"VNF placement algorithms to address the mono-and multi-tenant issues in edge and core networks","authors":"C. Morin, Géraldine Texier, C. Caillouet, Gilles Desmangles, Cao-Thanh Phan","doi":"10.1109/CloudNet47604.2019.9064108","DOIUrl":"https://doi.org/10.1109/CloudNet47604.2019.9064108","url":null,"abstract":"The Network Functions Virtualisation (NFV) concept offers network operators the ability to provide more scalable and less expensive services, free from the limitations inherent to hardware devices. However, in 5G networks, the functions must be deployed not only in large central data centers, but also in the edge. We propose an algorithm that solves the Virtual Network Function Chain Placement Problem allowing a fine management of these rare resources in order to respond to the greatest number of requests possible. Because networks can be divided into several entities belonging to different tenants who are reluctant to reveal their internal topologies, we propose a heuristic that allows the NFV orchestrator to place the function chains based only on an abstract view of the infrastructure network. We leverage this approach to address the complexity of the problem in large mono- or multi-tenant networks. We analyze the efficiency of our algorithm and heuristic with respect to a wide range of parameters and topologies.","PeriodicalId":340890,"journal":{"name":"2019 IEEE 8th International Conference on Cloud Networking (CloudNet)","volume":"67 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132046940","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-11-01DOI: 10.1109/CloudNet47604.2019.9064138
Naghmeh Dezhabad, S. Ganti, G. Shoja
Cloud providers aim to efficiently deliver diverse services on demand to users. Recently, they coined the idea of an auction-based market for their resources with the goal of increasing the total revenues. To address the challenge of scheduling and pricing, we build usage profiles for cloud workloads and predict future demands. In this paper, we first present a new methodology to categorize workloads according to their resource usage. We employ a modified hierarchical clustering algorithm that gives us three demand profiles for batch jobs designated as low, medium and high. After that, we extract the number of arrival requests per time for each group. The methodology presented here provides insights to cloud service providers in optimizing resource allocation and improving profits.
{"title":"Cloud Workload Characterization and Profiling for Resource Allocation","authors":"Naghmeh Dezhabad, S. Ganti, G. Shoja","doi":"10.1109/CloudNet47604.2019.9064138","DOIUrl":"https://doi.org/10.1109/CloudNet47604.2019.9064138","url":null,"abstract":"Cloud providers aim to efficiently deliver diverse services on demand to users. Recently, they coined the idea of an auction-based market for their resources with the goal of increasing the total revenues. To address the challenge of scheduling and pricing, we build usage profiles for cloud workloads and predict future demands. In this paper, we first present a new methodology to categorize workloads according to their resource usage. We employ a modified hierarchical clustering algorithm that gives us three demand profiles for batch jobs designated as low, medium and high. After that, we extract the number of arrival requests per time for each group. The methodology presented here provides insights to cloud service providers in optimizing resource allocation and improving profits.","PeriodicalId":340890,"journal":{"name":"2019 IEEE 8th International Conference on Cloud Networking (CloudNet)","volume":"66 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123031988","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-11-01DOI: 10.1109/CloudNet47604.2019.9064124
P. Valsamas, Panagiotis Papadimitriou, I. Sakellariou, S. Petridou, L. Mamatas, S. Clayman, F. Tusa, A. Galis
Network slicing is seen as a key enabler for meeting the diverse network service requirements, which stem from the transition to 5G. Furthermore, network slicing provides inherent support for multi-tenancy, enabling network providers to slice their infrastructure and resell it to a large number of tenants. Most existing work on slicing has been focused on certain mechanisms (e.g., slice embedding) and architecture specifications. As such, the performance and scalability with network slice instantiation has not been studied in depth. These aspects are even more critical in the case of slice deployments across multiple Points-of-Presence (PoP), since the various slice components should be stitched together for the end-to-end slice instantiation. In this paper, we present the design and prototype implementation of a network slicing architecture, based on which we perform a feasibility study of network slicing using multiple experimental infrastructures. Our prototype implementation supports all the required functionality for slice instantiation, such as resource discovery, slice embedding, resource provisioning, link setup, and inter-PoP slice segment stitching. Our experimental results corroborate the feasibility of multi-PoP network slicing. We further gain useful insights on slice instantiation performance and scalability.
{"title":"Multi-PoP Network Slice Deployment: A Feasibility Study","authors":"P. Valsamas, Panagiotis Papadimitriou, I. Sakellariou, S. Petridou, L. Mamatas, S. Clayman, F. Tusa, A. Galis","doi":"10.1109/CloudNet47604.2019.9064124","DOIUrl":"https://doi.org/10.1109/CloudNet47604.2019.9064124","url":null,"abstract":"Network slicing is seen as a key enabler for meeting the diverse network service requirements, which stem from the transition to 5G. Furthermore, network slicing provides inherent support for multi-tenancy, enabling network providers to slice their infrastructure and resell it to a large number of tenants. Most existing work on slicing has been focused on certain mechanisms (e.g., slice embedding) and architecture specifications. As such, the performance and scalability with network slice instantiation has not been studied in depth. These aspects are even more critical in the case of slice deployments across multiple Points-of-Presence (PoP), since the various slice components should be stitched together for the end-to-end slice instantiation. In this paper, we present the design and prototype implementation of a network slicing architecture, based on which we perform a feasibility study of network slicing using multiple experimental infrastructures. Our prototype implementation supports all the required functionality for slice instantiation, such as resource discovery, slice embedding, resource provisioning, link setup, and inter-PoP slice segment stitching. Our experimental results corroborate the feasibility of multi-PoP network slicing. We further gain useful insights on slice instantiation performance and scalability.","PeriodicalId":340890,"journal":{"name":"2019 IEEE 8th International Conference on Cloud Networking (CloudNet)","volume":"48 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128710839","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-11-01DOI: 10.1109/CloudNet47604.2019.9064090
Sara Shakeri, Niek van Noort, P. Grosso
Digital marketplaces (DMPs) are emerging as a framework for organizations to share their data. Security and support for multi-tenancy are the key features of DMPs. DMPs infrastructure can be built upon container-based networks in the cloud environments. However, there is not at the moment an in-depth analysis of the capability of container networks to support this mode of operation. In this paper, we evaluate the capability of Cilium and Calico, the two most popular container network techniques, in providing security (policy scalability) and handling the multi-tenancy requirements (pod scalability) of DMPs. We first measured the policy scalability in the network, and both Calico and Cilium scale well. However, by studying the pod scalability we determine there is around 50% throughput degradation in both technologies by increasing the number of pods from one to forty.
{"title":"Scalability of Container Overlays for Policy Enforcement in Digital Marketplaces","authors":"Sara Shakeri, Niek van Noort, P. Grosso","doi":"10.1109/CloudNet47604.2019.9064090","DOIUrl":"https://doi.org/10.1109/CloudNet47604.2019.9064090","url":null,"abstract":"Digital marketplaces (DMPs) are emerging as a framework for organizations to share their data. Security and support for multi-tenancy are the key features of DMPs. DMPs infrastructure can be built upon container-based networks in the cloud environments. However, there is not at the moment an in-depth analysis of the capability of container networks to support this mode of operation. In this paper, we evaluate the capability of Cilium and Calico, the two most popular container network techniques, in providing security (policy scalability) and handling the multi-tenancy requirements (pod scalability) of DMPs. We first measured the policy scalability in the network, and both Calico and Cilium scale well. However, by studying the pod scalability we determine there is around 50% throughput degradation in both technologies by increasing the number of pods from one to forty.","PeriodicalId":340890,"journal":{"name":"2019 IEEE 8th International Conference on Cloud Networking (CloudNet)","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132289075","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：