Landmark-based data location verification in the cloud: review of approaches and challenges

Journal of Cloud Computing

Advances, Systems and Applications

Table 1 Classification of landmark-assisted data verification approaches

	Correlation ^a	Landmark			Measurements collecting		PDP algorithm
	location / PIP@	Type	Distribution scale	Selection	Metrics	Probing protocol
Biswal2015 [27]	Yes	Active	USA	None^c	RTT, HC^d, BW^e	ICMP	None
Ries2011 [24]	Yes	Active	Worldwide	None^c	RTT	ICMP	None
Fotouhi2015 [10]	Yes	Active	Worldwide	Pre-verification	RTT	ICMP	None
Jaiswal2016 [22]	No	Active	USA	Pre-verification	RTT	ICMP, HTTP	None
Benson2011 [25]	No	Active	USA	None^c	RTT	HTTP	None
Gondree2013 [26]	No	Active	USA	None^c	RTT	HTTP	MAC-based
Watson2012 [21]	No	Active	USA and Europe	Pre-verification	RTT	HTTP	MAC-based
Eskandari2014 [23]	No	Passive^b	Worldwide	Pre-training	RTT	HTTP	None
	Machine learning
	Training coordination	Distance estimate		Location inference		Location granularity
Biswal2015 [27]	Centralized	None^f		Naive Bayes classification		County
Ries2011 [24]	Centralized	Virtual network coordinates		Instance-based Classification		Country
Fotouhi2015 [10]	Decentralized	Bestline		Multilateration		Area^g
Jaiswal2016 [22]	Decentralized	Distance to delay ratio		Multilateration		GPS coordinates
Benson2011 [25]	Centralized	Linear regression		Hierarchical clustering		City
Gondree2013 [26]	Decentralized	Bestline		Multilateration		Area^g
Watson2012 [21]	Decentralized	Linear regression		Multilateration		Area^g
Eskandari2014 [23]	Centralized	Polynomial regression		Multilateration		GPS coordinates

^aCorrelation between data location and proxy IP addresses
^bThis solution deploys only passive landmarks and measurements are performed by a virtual machine located on CSP
^cNone: all the landmarks in the initial set are used in the whole verification process
^dHC: number of hops on route
^eBW: bandwidth of route
^fThe classifier returns a location rather than a distance
^gThe size of the area depends on measurements