{"id":{"repo_id":"umkc","oai_identifier":"oai:mospace.umsystem.edu:10355/66836"},"canonical_url":"https://search.dev.ndltd.org/etd/umkc/oai:mospace.umsystem.edu:10355/66836","repository":{"repo_id":"umkc","name":"University of Missouri - Kansas City","base_url":"https://mospace.umsystem.edu/oai/request"},"display":{"title":"Cost Optimization and Load Balancing of Intra and Inter Data Center Networks to Facilitate Cloud Services","abstract":"For cloud enterprise customers that require services on demand, data centers (DC) must allocate and partition data center resources in a dynamic fashion. We consider the problem of allocating data center resources for cloud enterprise customers who require guaranteed services on demand. In particular, a request from an enterprise customer is mapped to a virtual network (VN) class that is allocated both bandwidth and compute resources by connecting it from an entry point of a data center to one or more hosts while there are multiple geographically distributed data centers to choose from. We take a dynamic traffic engineering approach over multiple time periods in which an energy aware resource reservation model is solved at each review point. In this dissertation, at first for the energy-aware resource reservation problem, we present a mixed-integer linear programming (MILP) formulation (for small-scale problems) and a heuristic approach (for large-scale problems). Our heuristic is fast for solving large-scale problems where the MILP problem becomes difficult to solve. Through a comprehensive set of studies, we found that a VN class with a low resource requirement has a low blocking even in heavy traffic, while the VN class with a high resource requirement faces a high service denial. Furthermore, the VN class having randomly distributed resource requirement has a high provisioning cost and blocking compared to the VN class having the same resource requirement for each request although the average resource requirement is same for both these VN classes. We also observe that our approach reduces the maximum energy consumption by about one-sixth at the low arrival rate to by about one-third at the highest arrival rate which also depends on how many different CPU frequency levels a server can run at. Allocation of resources in data centers needs to be done in a dynamic fashion for cloud enterprise customers who require virtualized reservation-oriented services on demand. Due to the spatial diversity of data centers, the cost of using different DCs also varies. In this dissertation, we then propose an allocation scheme to balance the load among these DCs with different cost to minimize the total provisioning cost in a dynamic environment while ensuring that the service level agreements (SLAs) are met. Compared to a benchmark scheme (where all requests are first sent to the cheapest data center), our scheme can decrease the proportional utilization from 24% (for heavy load) to 30% (for normal load) and achieve a significant balance in the cost incurred by individual DCs. Our scheme can also achieve 7.5% reduction in total provisioning cost under certain service level agreement (SLA) in exchange of low increment in blocking. Furthermore, we tested our scheme on 5 DCs to show that our allocation schemes follows the weighted cost proportionally. With the increasing dependency of cloud-based services, data centers have be come a popular platform to satisfy customers’ requests. Many large network providers now have their own geographically distributed DCs for cloud services, or have partner ships with third party DC providers to route customers’ demand. When end customers’ re quests arrive at a Point-of-Presence (PoP) of a large Internet Service Provider, the provider having DCs in multiple geo-locations needs to decide which DC should serve the request depending on the geo-distance, cost of resources in that DC, availability of the requested resource at that DC, and congestion in the path from the customers’ location to that DC. Therefore, an optimal connectivity scheme from the ingress PoP to egress DC is required among the PoPs and DCs to minimize the cost of establishing paths between a PoP and a DC while ensuring load balancing in both the link level and DC level. Considering these, we also present a novel mix-integer linear programming (MILP) model for this problem. We show the efficacy of our model through various performance metrics such as average and maximum link utilization, and average number of links used per path.","abstract_html":"For cloud enterprise customers that require services on demand, data centers (DC) must allocate and partition data center resources in a dynamic fashion. We consider the problem of allocating data center resources for cloud enterprise customers who require guaranteed services on demand. In particular, a request from an enterprise customer is mapped to a virtual network (VN) class that is allocated both bandwidth and compute resources by connecting it from an entry point of a data center to one or more hosts while there are multiple geographically distributed data centers to choose from. We take a dynamic traffic engineering approach over multiple time periods in which an energy aware resource reservation model is solved at each review point. In this dissertation, at first for the energy-aware resource reservation problem, we present a mixed-integer linear programming (MILP) formulation (for small-scale problems) and a heuristic approach (for large-scale problems). Our heuristic is fast for solving large-scale problems where the MILP problem becomes difficult to solve. Through a comprehensive set of studies, we found that a VN class with a low resource requirement has a low blocking even in heavy traffic, while the VN class with a high resource requirement faces a high service denial. Furthermore, the VN class having randomly distributed resource requirement has a high provisioning cost and blocking compared to the VN class having the same resource requirement for each request although the average resource requirement is same for both these VN classes. We also observe that our approach reduces the maximum energy consumption by about one-sixth at the low arrival rate to by about one-third at the highest arrival rate which also depends on how many different CPU frequency levels a server can run at. Allocation of resources in data centers needs to be done in a dynamic fashion for cloud enterprise customers who require virtualized reservation-oriented services on demand. Due to the spatial diversity of data centers, the cost of using different DCs also varies. In this dissertation, we then propose an allocation scheme to balance the load among these DCs with different cost to minimize the total provisioning cost in a dynamic environment while ensuring that the service level agreements (SLAs) are met. Compared to a benchmark scheme (where all requests are first sent to the cheapest data center), our scheme can decrease the proportional utilization from 24% (for heavy load) to 30% (for normal load) and achieve a significant balance in the cost incurred by individual DCs. Our scheme can also achieve 7.5% reduction in total provisioning cost under certain service level agreement (SLA) in exchange of low increment in blocking. Furthermore, we tested our scheme on 5 DCs to show that our allocation schemes follows the weighted cost proportionally. With the increasing dependency of cloud-based services, data centers have be come a popular platform to satisfy customers’ requests. Many large network providers now have their own geographically distributed DCs for cloud services, or have partner ships with third party DC providers to route customers’ demand. When end customers’ re quests arrive at a Point-of-Presence (PoP) of a large Internet Service Provider, the provider having DCs in multiple geo-locations needs to decide which DC should serve the request depending on the geo-distance, cost of resources in that DC, availability of the requested resource at that DC, and congestion in the path from the customers’ location to that DC. Therefore, an optimal connectivity scheme from the ingress PoP to egress DC is required among the PoPs and DCs to minimize the cost of establishing paths between a PoP and a DC while ensuring load balancing in both the link level and DC level. Considering these, we also present a novel mix-integer linear programming (MILP) model for this problem. We show the efficacy of our model through various performance metrics such as average and maximum link utilization, and average number of links used per path.","abstract_has_math":false,"creators":["Maswood, Mirza Mohd Shahriar"],"institution":"University of Missouri -- Kansas City","degree_name":"Ph.D.","degree_level":"Doctoral","degree_discipline":"Telecommunications and Computer Networking (UMKC)","degree_department":null,"school":null,"contributors":[],"advisors":["Medhi, Deepankar"],"committee_chairs":[],"committee_members":[],"year":2018,"date_issued":"2018","date_published":"2018","updated_at":"2026-07-24T05:18:34Z","subjects":[],"languages":["en_US"],"rights":[],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"https://hdl.handle.net/10355/66836","outbound_label":"Handle","outbound_source":"dc:identifier.uri"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.advisor","label":"Advisor","values":["Medhi, Deepankar"]},{"key":"dc:creator","label":"Author","values":["Maswood, Mirza Mohd Shahriar"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.accessioned","label":"Dc Date Accessioned","values":["2018-12-20T15:20:32Z"]},{"key":"dc:date.available","label":"Dc Date Available","values":["2018-12-20T15:20:32Z"]},{"key":"dc:date.issued","label":"Date","values":["2018"]},{"key":"dc:publisher","label":"Institution","values":["University of Missouri -- Kansas City"]},{"key":"dc:type","label":"Dc Type","values":["Thesis"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Telecommunications and Computer Networking (UMKC)","Electrical and Computer Engineering (UMKC)"]},{"key":"thesis:degree_level","label":"Degree Level","values":["Doctoral"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Ph.D."]},{"key":"thesis:institution_name","label":"Thesis Institution Name","values":["University of Missouri--Kansas City"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language.iso","label":"Language (ISO)","values":["en_US"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.uri","label":"Identifier URI","values":["https://hdl.handle.net/10355/66836"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["Title from PDF of title page viewed January 3, 2019","Dissertation advisor: Deep Medhi","Vita","Includes bibliographical references (pages 127-137}","Thesis (Ph.D.)--School of Computing and Engineering. University of Missouri--Kansas City, 2018"]},{"key":"dc:description.abstract","label":"Abstract","values":["For cloud enterprise customers that require services on demand, data centers (DC) must allocate and partition data center resources in a dynamic fashion. We consider the problem of allocating data center resources for cloud enterprise customers who require guaranteed services on demand. In particular, a request from an enterprise customer is mapped to a virtual network (VN) class that is allocated both bandwidth and compute resources by connecting it from an entry point of a data center to one or more hosts while there are multiple geographically distributed data centers to choose from. We take a dynamic traffic engineering approach over multiple time periods in which an energy aware resource reservation model is solved at each review point. In this dissertation, at first for the energy-aware resource reservation problem, we present a mixed-integer linear programming (MILP) formulation (for small-scale problems) and a heuristic approach (for large-scale problems). Our heuristic is fast for solving large-scale problems where the MILP problem becomes difficult to solve. Through a comprehensive set of studies, we found that a VN class with a low resource requirement has a low blocking even in heavy traffic, while the VN class with a high resource requirement faces a high service denial. Furthermore, the VN class having randomly distributed resource requirement has a high provisioning cost and blocking compared to the VN class having the same resource requirement for each request although the average resource requirement is same for both these VN classes. We also observe that our approach reduces the maximum energy consumption by about one-sixth at the low arrival rate to by about one-third at the highest arrival rate which also depends on how many different CPU frequency levels a server can run at. Allocation of resources in data centers needs to be done in a dynamic fashion for cloud enterprise customers who require virtualized reservation-oriented services on demand. Due to the spatial diversity of data centers, the cost of using different DCs also varies. In this dissertation, we then propose an allocation scheme to balance the load among these DCs with different cost to minimize the total provisioning cost in a dynamic environment while ensuring that the service level agreements (SLAs) are met. Compared to a benchmark scheme (where all requests are first sent to the cheapest data center), our scheme can decrease the proportional utilization from 24% (for heavy load) to 30% (for normal load) and achieve a significant balance in the cost incurred by individual DCs. Our scheme can also achieve 7.5% reduction in total provisioning cost under certain service level agreement (SLA) in exchange of low increment in blocking. Furthermore, we tested our scheme on 5 DCs to show that our allocation schemes follows the weighted cost proportionally. With the increasing dependency of cloud-based services, data centers have be come a popular platform to satisfy customers’ requests. Many large network providers now have their own geographically distributed DCs for cloud services, or have partner ships with third party DC providers to route customers’ demand. When end customers’ re quests arrive at a Point-of-Presence (PoP) of a large Internet Service Provider, the provider having DCs in multiple geo-locations needs to decide which DC should serve the request depending on the geo-distance, cost of resources in that DC, availability of the requested resource at that DC, and congestion in the path from the customers’ location to that DC. Therefore, an optimal connectivity scheme from the ingress PoP to egress DC is required among the PoPs and DCs to minimize the cost of establishing paths between a PoP and a DC while ensuring load balancing in both the link level and DC level. Considering these, we also present a novel mix-integer linear programming (MILP) model for this problem. We show the efficacy of our model through various performance metrics such as average and maximum link utilization, and average number of links used per path."]},{"key":"dc:title","label":"Title","values":["Cost Optimization and Load Balancing of Intra and Inter Data Center Networks to Facilitate Cloud Services"]}]}],"canonical_facts":{"dc:contributor.advisor":["Medhi, Deepankar"],"dc:creator":["Maswood, Mirza Mohd Shahriar"],"dc:date.accessioned":["2018-12-20T15:20:32Z"],"dc:date.available":["2018-12-20T15:20:32Z"],"dc:date.issued":["2018"],"dc:description":["Title from PDF of title page viewed January 3, 2019","Dissertation advisor: Deep Medhi","Vita","Includes bibliographical references (pages 127-137}","Thesis (Ph.D.)--School of Computing and Engineering. University of Missouri--Kansas City, 2018"],"dc:description.abstract":["For cloud enterprise customers that require services on demand, data centers (DC) must allocate and partition data center resources in a dynamic fashion. We consider the problem of allocating data center resources for cloud enterprise customers who require guaranteed services on demand. In particular, a request from an enterprise customer is mapped to a virtual network (VN) class that is allocated both bandwidth and compute resources by connecting it from an entry point of a data center to one or more hosts while there are multiple geographically distributed data centers to choose from. We take a dynamic traffic engineering approach over multiple time periods in which an energy aware resource reservation model is solved at each review point. In this dissertation, at first for the energy-aware resource reservation problem, we present a mixed-integer linear programming (MILP) formulation (for small-scale problems) and a heuristic approach (for large-scale problems). Our heuristic is fast for solving large-scale problems where the MILP problem becomes difficult to solve. Through a comprehensive set of studies, we found that a VN class with a low resource requirement has a low blocking even in heavy traffic, while the VN class with a high resource requirement faces a high service denial. Furthermore, the VN class having randomly distributed resource requirement has a high provisioning cost and blocking compared to the VN class having the same resource requirement for each request although the average resource requirement is same for both these VN classes. We also observe that our approach reduces the maximum energy consumption by about one-sixth at the low arrival rate to by about one-third at the highest arrival rate which also depends on how many different CPU frequency levels a server can run at. Allocation of resources in data centers needs to be done in a dynamic fashion for cloud enterprise customers who require virtualized reservation-oriented services on demand. Due to the spatial diversity of data centers, the cost of using different DCs also varies. In this dissertation, we then propose an allocation scheme to balance the load among these DCs with different cost to minimize the total provisioning cost in a dynamic environment while ensuring that the service level agreements (SLAs) are met. Compared to a benchmark scheme (where all requests are first sent to the cheapest data center), our scheme can decrease the proportional utilization from 24% (for heavy load) to 30% (for normal load) and achieve a significant balance in the cost incurred by individual DCs. Our scheme can also achieve 7.5% reduction in total provisioning cost under certain service level agreement (SLA) in exchange of low increment in blocking. Furthermore, we tested our scheme on 5 DCs to show that our allocation schemes follows the weighted cost proportionally. With the increasing dependency of cloud-based services, data centers have be come a popular platform to satisfy customers’ requests. Many large network providers now have their own geographically distributed DCs for cloud services, or have partner ships with third party DC providers to route customers’ demand. When end customers’ re quests arrive at a Point-of-Presence (PoP) of a large Internet Service Provider, the provider having DCs in multiple geo-locations needs to decide which DC should serve the request depending on the geo-distance, cost of resources in that DC, availability of the requested resource at that DC, and congestion in the path from the customers’ location to that DC. Therefore, an optimal connectivity scheme from the ingress PoP to egress DC is required among the PoPs and DCs to minimize the cost of establishing paths between a PoP and a DC while ensuring load balancing in both the link level and DC level. Considering these, we also present a novel mix-integer linear programming (MILP) model for this problem. We show the efficacy of our model through various performance metrics such as average and maximum link utilization, and average number of links used per path."],"dc:identifier.uri":["https://hdl.handle.net/10355/66836"],"dc:language.iso":["en_US"],"dc:publisher":["University of Missouri -- Kansas City"],"dc:title":["Cost Optimization and Load Balancing of Intra and Inter Data Center Networks to Facilitate Cloud Services"],"dc:type":["Thesis"],"thesis:degree_discipline":["Telecommunications and Computer Networking (UMKC)","Electrical and Computer Engineering (UMKC)"],"thesis:degree_level":["Doctoral"],"thesis:degree_name":["Ph.D."],"thesis:institution_name":["University of Missouri--Kansas City"]},"updated_at":"2026-07-24T05:18:34Z"}