DocumentCode
725339
Title
Shuttle: Intrusion Recovery for PaaS
Author
Nascimento, Dario ; Correia, Miguel
Author_Institution
Inst. Super. Tocnico, Univ. de Lisboa, Lisbon, Portugal
fYear
2015
fDate
June 29 2015-July 2 2015
Firstpage
653
Lastpage
663
Abstract
The number of applications being deployed using the Platform as a Service (PaaS) cloud computing model is increasing. Despite the security controls implemented by cloud service providers, we expect intrusions to strike such applications. We present Shuttle, a novel intrusion recovery service. Shuttle recovers from intrusions in applications deployed in PaaS platforms. Our approach allows undoing changes to the state of PaaS applications due to intrusions, without loosing the effect of legitimate operations performed after the intrusions take place. We combine a record-and-replay approach with the elasticity provided by cloud offerings to recover applications deployed on various instances and backed by distributed databases. The service loads a database snapshot taken before the intrusion and replays subsequent requests, as much in parallel as possible, while continuing to execute incoming requests. We present an experimental evaluation of Shuttle on Amazon Web Services. We show Shuttle can replay 1 million requests in 10 minutes and that it can duplicate the number of requests replayed per second by increasing the number of application servers from 1 to 3.
Keywords
Web services; cloud computing; distributed databases; security of data; Amazon Web services; PaaS platforms; Shuttle; application servers; cloud computing model; cloud service providers; database snapshot; distributed databases; intrusion recovery service; platform as a service; record-and-replay approach; security controls; time 10 min; Computational modeling; Distributed databases; Elasticity; Security; Servers; Software; Cloud Computing; Dependability; Distributed Database Systems; Intrusion Recovery; Intrusion Tolerance; Platform as a Service;
fLanguage
English
Publisher
ieee
Conference_Titel
Distributed Computing Systems (ICDCS), 2015 IEEE 35th International Conference on
Conference_Location
Columbus, OH
ISSN
1063-6927
Type
conf
DOI
10.1109/ICDCS.2015.72
Filename
7164950
Link To Document