Show simple item record

dc.identifier.urihttp://hdl.handle.net/11401/77311
dc.description.sponsorshipThis work is sponsored by the Stony Brook University Graduate School in compliance with the requirements for completion of degree.en_US
dc.formatMonograph
dc.format.mediumElectronic Resourceen_US
dc.language.isoen_US
dc.publisherThe Graduate School, Stony Brook University: Stony Brook, NY.
dc.typeDissertation
dcterms.abstractNetworks today rely on network functions or middleboxes (e.g., firewalls, WAN optimizers) to provide critical performance, security, and policy compliance capabilities. However, today the management of these middleboxes is hard. First, these middleboxes are implemented as dedicated hardware appliances, making it difficult to dynamically scale resources. Second, operators need to carefully plan the network topology, manually set up rules to route traffic through the desired sequence of middleboxes, and implement safeguards for correct operation in the presence of failures and overload. We can overhaul today’s network infrastructure to address these problems by introducing flexibility in routing and the implementation of these middleboxes. However, from network operator’s perspective a key question is whether we can address these problems in a minimally disruptive manner, e.g., which require minimal changes to existing middlebox implementations and routing mechanisms. In this thesis, I describe two case studies for introducing more flexibility in middlebox management with minimal changes to existing middlebox implementations and routing mechanisms. In the first part of the thesis, I describe SIMPLE, a Software-Defined Networking (SDN) based efficient middlebox traffic steering solution which works with existing middlebox implementations and uses existing SDN APIs. In SIMPLE, I address algorithmic and system design challenges to demonstrate the feasibility of using SDN to simplify middlebox traffic steering. In the second part of the thesis, I describe KLEIN a cellular core re-design that uses Network Function Virtualization (NFV) and smart resource management, stays within the confines of current cellular standards and uses legacy routing in the core network. I address key challenges w.r.t. scalability, responsiveness and in realizing KLEIN via backwards-compatible orchestration mechanisms.
dcterms.abstractNetworks today rely on network functions or middleboxes (e.g., firewalls, WAN optimizers) to provide critical performance, security, and policy compliance capabilities. However, today the management of these middleboxes is hard. First, these middleboxes are implemented as dedicated hardware appliances, making it difficult to dynamically scale resources. Second, operators need to carefully plan the network topology, manually set up rules to route traffic through the desired sequence of middleboxes, and implement safeguards for correct operation in the presence of failures and overload. We can overhaul today’s network infrastructure to address these problems by introducing flexibility in routing and the implementation of these middleboxes. However, from network operator’s perspective a key question is whether we can address these problems in a minimally disruptive manner, e.g., which require minimal changes to existing middlebox implementations and routing mechanisms. In this thesis, I describe two case studies for introducing more flexibility in middlebox management with minimal changes to existing middlebox implementations and routing mechanisms. In the first part of the thesis, I describe SIMPLE, a Software-Defined Networking (SDN) based efficient middlebox traffic steering solution which works with existing middlebox implementations and uses existing SDN APIs. In SIMPLE, I address algorithmic and system design challenges to demonstrate the feasibility of using SDN to simplify middlebox traffic steering. In the second part of the thesis, I describe KLEIN a cellular core re-design that uses Network Function Virtualization (NFV) and smart resource management, stays within the confines of current cellular standards and uses legacy routing in the core network. I address key challenges w.r.t. scalability, responsiveness and in realizing KLEIN via backwards-compatible orchestration mechanisms.
dcterms.available2017-09-20T16:52:28Z
dcterms.contributorDas, Samiren_US
dcterms.contributorBalasubramanian, Arunaen_US
dcterms.contributorGill, Phillipaen_US
dcterms.contributorSekar, Vyasen_US
dcterms.contributorGopalakrishnan, Vijay.en_US
dcterms.creatorQazi, Zafar Ayyub
dcterms.dateAccepted2017-09-20T16:52:28Z
dcterms.dateSubmitted2017-09-20T16:52:28Z
dcterms.descriptionDepartment of Computer Science.en_US
dcterms.extent138 pg.en_US
dcterms.formatApplication/PDFen_US
dcterms.formatMonograph
dcterms.identifierhttp://hdl.handle.net/11401/77311
dcterms.issued2015-12-01
dcterms.languageen_US
dcterms.provenanceMade available in DSpace on 2017-09-20T16:52:28Z (GMT). No. of bitstreams: 1 Qazi_grad.sunysb_0771E_12670.pdf: 3939722 bytes, checksum: 1bf5882cc74ee986038451c985c9f8fe (MD5) Previous issue date: 1en
dcterms.publisherThe Graduate School, Stony Brook University: Stony Brook, NY.
dcterms.subjectComputer science
dcterms.titleMinimally Disruptive Management Frameworks for Network Functions
dcterms.typeDissertation


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record