Application Security Review - VyprVPN
VyprVPN No Log AssessmentFinal ReportPrepared for: Michael DouglassChief ScientistGolden Frog2500 Bee Cave RoadBuilding 1, Suite 400Austin, TX 78746-6943United States11/9/2018
2018 Leviathan Security Group Incorporated.All Rights Reserved.DisclaimerThis document contains information, which isNo trademark, copyright, or patent licenses areprotected by copyright and pre-existing non-expressly or implicitly granted (herein) with thisdisclosure agreement between Leviathananalysis, report, or white paper.Security and the company identified as“PdF ”hAll brand names and product names used in thisgdocument are trademarks, registeredNo part of this document may be photocopied,trademarks, or trade names of their respectivereproduced, or translated to another languageholders. Leviathan Security Group is notwithout the prior written and documentedassociated with any other vendors or productsconsent of Leviathan Security Group and thementioned in this document.y df d“PdF ”hpage.Version:Final Report with Retest ResultsPrepared for:Golden FrogDate:11/9/2018Prepared by:JV, RyO’N,D CC w, and Mark StriblingConfidentiality NoticeThis document contains information confidential and proprietary to Leviathan Security Group and GoldenFrog. The information may not be used, disclosed or reproduced without the prior written authorization ofeither party and those so authorized may only use the information for the purpose of evaluationconsistent with authorization. Reproduction of any section of this document must include this notice.
Table of ContentsTable of Contents . 3Executive Summary . 4Observations . 4Recommendations . 4Findings Index . 5Log Finding Classification . 6Threat Assessment and Scoping . 7Project Components and Descriptions . 9Observations & Analysis . 11Cruncher/Logs and Splunk . 11VPN Termination . 13DAPI . 17Control Panel API. 19RADIUS Proxy and Server . 21User Client Applications . 23Lease DB . 25Appendix A – Technical Services . 26Appendix B – Risk and Advisory Services . 27Appendix C – Leviathan Project Team . 28CONFIDENTIAL INFORMATION / LIMIT DISTRIBUTION3
Executive SummaryGolden Frog engaged Leviathan Security Group in September 2018 to evaluate updates to its VyprVPNproduct with the goal of providing “N L g” VPNuSfy, Lhwedthe platform to ensure that no Personally-Identifiable Information (PII) about customers with respect totheir use of the service are logged by Golden Frog except as required for business operations.Golden Frog seeks to reduce the exposure of their clients by providing VPN service with an explicitguarantee that it does not log connection details. It designed the various clients to report informationabout connections by default, but that connection data is anonymized and can be turned off by userconfiguration. Connections are logged during authentication, but logs that could identify users are keptonly for a short time. By using open source or widely-used applications for server and client components,they have reduced the risk of unintentionally adding a weakness into the software themselves.ObservationsThe VPN clients produce no identifying logs w h u h u’There were a limited number ofidentifying logs on the VPN server which appeared result from inadvertent configuration mistakes.Specifically, we found that logs for the kill switch API contained user IP addresses.We considered a wide range of attackers that might be interested in deanonymizing VPN users. The mostcommon ones would be an ISP attempting to learn about the browsing habits of a user, or contentproducers seeking to restrict activity to a geographic region. The most powerful attackers would be nationstates that may wish to determine the source of a leak. While VPNs should not be considered an effectivedefense against highly-sophisticated attackers, they can reduce thosek’ options from surreptitiousgathering of intelligence to issuing a writ to force a company to spy on their users.We examined all components of the project according to the threat assessment described below. Whilevigilance against logging is necessary to complete thehdgfugwkfG dg “N L g”, w fhhF g’ implementation. The projectrevealed a limited number of issues that Golden Frog quickly fixed. As a result, it can provide VyprVPNusers with the assurance that the company is not logging their VPN activity.RecommendationsGolden Frog worked to remediate all no-log-related findings concurrently with the assessment. Once ithad completed this, we performed a retest and verified that all of the fixes were effective.In the long term, we recommend that Golden Frog continue to monitor their servers for regressions. Thiscan be done manually as part of maintenance or automatically by creating an alert that triggers on testdata being recorded in logs. Since Golden Frog already has test accounts, they could use these to greplogs for IP, username, and user ID. While this will not find different encodings or data that could becombined to deanonymize a user, it avoids the most common case where a configuration change resultsin a regression to the default behavior.CONFIDENTIAL INFORMATION / LIMIT DISTRIBUTION4
Findings IndexThis section represents a quick view into the findings discovered in this tive information in openvpn.logVPN83898FixedHighSensitive information in ppp-VPN83900FixedCruncher/Logs and83869Fixedconnect.logHighSplunk servers contain sensitive logdata (usernames and IPs)SplunkMediumSensitive information in dmesgVPN83899FixedMediumSensitive information in killswitch.logVPN83901FixedMediumURL path in DAPI logsDAPI83868FixedCONFIDENTIAL INFORMATION / LIMIT DISTRIBUTION5
Log Finding ClassificationImpactWhen we find logging of PII, we assign it one of five categories of severity,essentially describing the potential impact if it was abused:Informational only – W f u ddbut it could create one in the futufhdh’ugdhY u’,b bywant to fix it.Low – The log might allow an attacker to gain information that could besensitive H w,d’w direct access to data or resources.Medium – The log may result in an activity being connected to a user.High – The log reveals all information or activity of a user.Critical – The log reveals all information about all users.Skill Level toWhen we find logging of PII, we also assess what authority a user must have toExploitgain access to it:Simple – Only a minimal understanding of the underlying technology isrequired. Tools and techniques for accessing it can be easily found on theInternet.Moderate – The data is not available to all users, but it is available to employees.It is available to attackers who are able to exploit a vulnerability.Advanced – Near-complete and superior understanding of the technologyinvolved is required. Direct interaction with the victim or target may also berequired.ImpactRating(Weight)Authority Level to Access Rating (Weight)SeverityCritical (4)4812CriticalHigh (3)369High7-9Medium (2)246Medium4-6Low (1)123Low1-3Advanced (1)Moderate (2)Simple (3)CONFIDENTIAL INFORMATION / LIMIT DISTRIBUTION10-126
Threat Assessment and ScopingThis engagement was notably not a security evaluation, but rather focused on privacy. The goal was todetermine whether VyprVPN was logging user activities. Golden Frog explicitly seeks to avoid logging anysensitive activity related to its users use of the VyprVPNAu h, hgu“h”u dconsideration is the propensity of the specific software components used by VyprVPN to perform loggingwhich can associate a user with a specific VyprVPN session or to any activity during the session, andwhether Golden Frog was successful in removing or mitigating that logging.The system looks like this. We collapsed complex components into a single box in the figure when wedetermined that they comprise a single security principal1. The components have been colorized withrespect to the level of threat presented to that component. Red indicates a major threat in that thehfu dh u’uydudgOrange indicates a minor threat inthat the component has some user data that could be concerning and could be logged. Normal black textindicates that the component is not at significant risk of logging user activity, because the componentdoes not have the data. Blue indicates components that would be a security risk but are out of scope for aspecified reason.Control entContent d BalancerVPN onLeaseDBRADIUSCustomerDBSplunkBIA set of components that all have exactly the same set of access and privilegeshttps://en.wikipedia.org/wiki/Principal (computer security)1CONFIDENTIAL INFORMATION / LIMIT DISTRIBUTION7
User: This is the customer/user of the VyprVPN service Th uF g’, bu hfwd dh u’uby G ddF g G dfG dF g ffclient applications for Windows, MacOS, iOS, and Android. The client software is a moderate riskof logging, in that it does actually log a great deal of information, but it does not export thatinformation unless the user consents. In several cases, it is not possible for Golden Frog toprevent the logging, because the OS platform does not provide options to disable logging. Thesoftware should be inspected to assure that logs are only exported with user consent. Load Balancers: Users connect to the Load Balancer which uses standard networking techniquesto hand the connection off to the VPN Termination server Th u’h LdBalancer is thus only momentary, but none the less the Load Balancer witnesses a user connectingto VyprVPN, and so is a moderate risk of logging. VPN Termination: The client communicates with the VPN Termination server via one of 4protocols (OpenVPN2, IPSec3, PPTP4, and L2TP5). The user first authenticates to the VPNtermination point, which then negotiates a VPNw h h u’fwThVPN Termination node is a high risk of logging, because upon connection and throughout the,fuf h u’ dydThfwwas inspected toassure that it does not log these activities. Special consideration was made to differentiatebetween Connection State information which lives for the duration of a connection vs. Loggedinformation kept after the termination of a session. DAPI: In the course of operations, the client software calls various APIs offered by Golden Frogand is served by the DAPI node. DAPI is a high risk of logging, because upon calling its APIs, it isfuf h u’ dydThfwwas inspected to assure that itdoes not log these activities in association with the identity of a user. RADIUS Proxy, RADIUS: The VyprVPN Termination servers authenticate VyprVPN customersusing the RADIUS protocol. The VPN Termination node use the RADIUS Proxy/Server running onthe VyprVPN Termination server itself to load-balance the RADIUS request to the RADIUS serversin the site. The RADIUS proxy and server are a high risk of logging, because upon connecting toVy VPN,fuf h u’ dy, h CIP, h LIP,d hwhich the user connected. The software was inspected to assure that it does not logauthentication events in a way which allows the association of a user with a session. Note that notlogging authentication attempts and successes is itself a security threat, and so Golden Frog has avery challenging conflict of competing interests here. Lease DB: RADIUS uses the Lease DB to manage user network leases. The Lease DB is amoderate risk fgg g, buh u’gd IP ddfw kmanagement purposes for the duration of the connection. The Lease DB was inspected to assurethat the data it stores is the minimum required to achieve functional objectives, and that the datais not persisted any longer than en.wikipedia.org/wiki/Point-to-Point Tunneling Protocol5https://en.wikipedia.org/wiki/Layer 2 Tunneling Protocol23CONFIDENTIAL INFORMATION / LIMIT DISTRIBUTION8
Customer DB: RADIUS and DAPI both use the Customer DB to manage customer status, such ascredentials and billing status. The Customer DB is a not at risk of logging, in that it is a read-onlydatabase of customers. Customer information is a privacy concern, but Golden Frog is onlypromising not to log customer activity; the fact that a person is a Golden Frog customer isnecessarily logged for authentication and payment purposes. Control Panel API: Golden Frog uses two private API endpoints to allow customers to disconnecttheir connections via a central web control panel. The endpoints access the lease database and, ifnecessary, make a request to the VPN process itself to terminate a connection. Requests originatefrom the Golden Frog website server in Zurich. These requests and the responses may include theu’ dyd hIP ddAu , hdhigh risk oflogging. The software was inspected to verify that neither the responses nor the requests log anyof this information. The serv/vyprapi systems takes the request and then contacts the ControlPanel API on the proper vpn/vyprnode system to do the actual work. DNS: VyprVPN ffDNS,hu’ DNS qududfhuser. The VyprVPN DNS server in turn queries the Internet to resolve DNS queries, sourcing theDNS request from the appropriate geographic node. The DNS resolver works in two stages: Alocal DNS forwarder runs in each site. Customer DNS requests go to this forwarder first and comefrom the Lease IP of the VPN connection. The forwarder then sends the DNS request to an incountry VM; the in-country VM has zero knowledge of the user because all requests come fromthe DNS forwarder IP address. The DNS resolver is not at risk of logging, because it never haspossession of specific user identities causing the requests that it serves. Transparent Content Proxy: The Transparent Content Proxy is not at risk of logging, because itnever has specific customer identities. LogicMonitor: This is a system monitoring log server, used by Golden Frog to manageperformance and resource allocation. It does not handle specific-user data, instead loggingactivities like CPU and disk utilization, availability, and server timings. Thus, LogicMonitor is not atrisk of logging because it does not have any user-specific data. Cruncher/Logs, Splunk: These nodes are used for detailed system logging, and thus well couldbe storing specific user activity events. Sanitizing logs of user-specific activity is notoriouslydifficult6, and so they are a high risk of logging.Project Components and DescriptionsFrom the threat assessment above, these are the project components to be evaluated and how weassessed each one.CRUNCHER/LOGS, SPLUNKInspect to assure that it is not logging particular user activities, andthat user activities cannot be inferred from the logs it does record6https://en.wikipedia.org/wiki/Netflix Prize#Privacy concernsCONFIDENTIAL INFORMATION / LIMIT DISTRIBUTION9
VPN TERMINATION &Inspect to assure that it is not logging user connectionsCONTROL PANEL APIDAPIInspect to assure that it is not logging user API callsCONTROL PANEL APIInspect to assure that it is not logging user API callsRADIUS PROXY AND SERVERInspect to assure that it is not logging user authenticationUSER CLIENT APPLICATIONInspect to assure that the logs it stores are not exported without userconsentLEASE DBInspect to assure that it is only storing the minimum user informationnecessary for functional objectives, and not persist longer thannecessaryLOAD BALANCERSInspect to assure that the load balancers do not log connections asthey handle user requests and pass them along to actual serversCONFIDENTIAL INFORMATION / LIMIT DISTRIBUTION10
Observations & AnalysisThe following sections describe the results of our evaluation of each component.Cruncher/Logs and SplunkVyprVPN uses Cruncher/Logs to send event logs to Splunk7. Splunk is a product that collects data andlog-files from remote machines and presents the data via a Web interface that allows for queries andsearches. Splunk can also be used to filter data found in logs, and create complex pattern matchingstatements for omitting, mutating, removing or augmenting the data that it is collecting. The Splunkforwarders were running on two servers cruncher1 and logs1.Threat AnalysisSome log events are important to Golden Frog and harmless to users, such as when a particular disk drivefuhfd Ohw uduf h “g”u, u hfailed login attempts by legitimate users mistyping their password.MethodologyGolden Frog did not provide source code for Cruncher nor Splunk as they are 3rd party products. Insteadwe were able to observe configuration files on the Splunk servers. We also analyzed their respective Linuxenvironment to look for sensitive logging data. We searched for unsafe permissions on log data and corefile ulimit settings. We also used simple pattern matching searches against the log files in variouslocations.Observed TrendsThe Splunk servers both have coredumps explicitly disabled; ulimit -c is set to 0. Coredumps fromSplunk would have all the recent logs stored in the process memory of the data segment, heap and stackand would thus be logged within the core file itself.The Splunk configuration itself is fairly complex and we did not ascertain every nuance of itsconfiguration. However, we did observe that log-files appeared to be sanitized by the time they reach theSplunk server. That said, there were several locations on the 'logs1' server where log-files containedsensitive user data including usernames and IPs (83869). The retest confirmed that Golden Frog addressedthis issue.7https://www.splunk.com/CONFIDENTIAL INFORMATION / LIMIT DISTRIBUTION11
FindingsSPLUNK SERVERS CONTAIN SENSITIVE LOG DATA (USERNAMES AND IPS)ID83869ComponentCruncher/Logs and SplunkSeverityHighImpact / Skill onThe application does not take measures to sanitize sensitive information out of the log output. In thefollowing log-files we found usernames in the form of username@email.com, along with two GoldenFrog IP addresses; the called ip and the nas gOct 9 06:11:28 vyprnode116.ams1.goldenfrog.com ip-down.local: clientdisconnect success, username email address protocol pptp pppdevice ppp6nas ip 10.102.113.116 framed ip 31.6.27.110Oct 9 03:35:07 vyprnode13.ams1.goldenfrog.com ip-up.local: client-connectsuccess, username email address protocol l2tp pppdevice ppp5nas ip 10.102.113.13 framed ip 128.90.24.87Oct 9 03:35:03 vyprnode96.ams1.goldenfrog.com ip-down.local: clientdisconnect start, username email address waa protocol pptp pppdevice ppp2nas ip 10.102.113.96 framed ip 31.6.32.86RecommendationMake sure that all logs are properly sanitized so that there is no usernames or other sensitiveinformation being logged.CONFIDENTIAL INFORMATION / LIMIT DISTRIBUTION12
VPN TerminationThis node is the termination point for VPN services. It serves connections via OpenVPN, IPsec, PPTP, andL2TP. It also is the front face of user authentication, validating that a requesting user is actually a VyprVPNcustomer.Threat AnalysisThe VPN termination node has complete awareness of all user activities: Initial connection Authentication All network traffic DNS requests TerminationIt is vital that VPN not log this user activity.MethodologyWe viewed the components of the system and reviewed any modifications necessary for commercial use.We then logged into two production servers provided for testing and checked their configuration. Thenwe checked running processes to determine what files might be in use. We looked at each log filemanually and grepped for usernames (which have a common format), user IDs, and IP addresses (whichare regular) being logged. We checked the operating system for any logging mechanism that could beused to track user activities.Observed TrendsThe design of the system uses off-the-shelf servers (StrongSwan, OpenVPN, OpenL2TPd, and pptpd) withmodifications to fit into this commercial use case. Because these servers are off-the-shelf, they are proneto logging information that users would prefer to be private. These behaviors need to be modified to getthe desired result. In the case of OpenVPN, PPTP, and L2TP, we found that logging of usernames occurred(83898 and 83900). This can be explained by the fact that configuration management might not besufficiently rigorous. We also found that dmesg was logging IP addresses (83899).The retest confirmed that Golden Frog successfully addressed all issues in VPN termination.CONFIDENTIAL INFORMATION / LIMIT DISTRIBUTION13
FindingsSENSITIVE INFORMATION IN OPENVPN.LOGID83898ComponentVPNSeverityHighImpact / Skill nfrog.com logStatusFixedObservationThe application does not take measures to sanitize sensitive information out of the log output.The log contains email address and IP address of the user.Oct 8 03:27:48 vpn1.tor.goldenfrog.com openvpn: Mon Oct 8 03:27:48 2018 email address waa/ ip address :63655 WARNING: 'link-mtu' is usedinconsistently, local 'link-mtu 1570', remote 'link-mtu 1602'Oct 9 03:20:06 vyprnode1.ams1.goldenfrog.com openvpn: Tue Oct 9 03:20:062018 email address maa/ ip address :1194 WARNING: 'link-mtu' is usedinconsistently, local 'link-mtu 1570', remote 'link-mtu 1602'RecommendationCheck all instance of calling a log writer to make sure that sensitive information such as username andIP are not being logged.CONFIDENTIAL INFORMATION / LIMIT DISTRIBUTION14
SENSITIVE INFORMATION IN ct / Skill .goldenfrog.com The application does not take measures to sanitize sensitive information out of the log outputThe ppp-connect.log logs usernames and IP addresses.Oct 8 06:42:36 vyprnode1.ams1.goldenfrog.com ip-up.local: client-connectstart, username email address protocol pptp pppdevice ppp3nas ip 10.102.113.1 framed ip IP address RecommendationCheck all instance of calling a log writer to make sure that sensitive information such as usernames andIP addresses are not being logged.CONFIDENTIAL INFORMATION / LIMIT DISTRIBUTION15
SENSITIVE INFORMATION IN DMESGID83899ComponentVPNSeverityMediumImpact / Skill denfrog.com dmesgStatusFixedObservationThe application does not take measures to sanitize sensitive information out of the log output.The kernel logs IP address and port of systems that send packets with a bad checksum. This can beused to identify users because UDP packets will have a random mutation over time. Users of the VPNwill eventually send a packet with a bad checksum.dmesg[77274.828361] UDP: bad checksum. From IP address :51601 to IPaddress :16802 ulen 108RecommendationCheck all instance of calling a log writer to make sure that sensitive information such as IP address andusername are not being logged.CONFIDENTIAL INFORMATION / LIMIT DISTRIBUTION16
DAPIDAPI is an API server, used by the VyprVPNfwgh u’Threat AnalysisThe DAPI node has awareness of a subset of the user activities exposed to the VPN node Initial connection Authentication TerminationIt is important that VPN not log this user activity.MethodologyWe started the test by reading the source code to find what the system did and where logging couldoccur. We requested data from the server and recorded the responses. We were provided a trial premiumaccount, so we used the credentials to make authenticated requests to the server. We logged into theweb server and checked for configuration and logs in the place where Apache is configured to store logs.We additionally looked in the syslog and other system logs for evidence of users. We looked at processesrunning to ensure that the system did not have any software that would create logs besides Apache.Observed TrendsThe DAPI is a web service built on Flask with authentication of certain necessary endpoints. It is used forunauthenticated calls to /vyprvpn/locations which provides users with a list of servers they can connect to.It is also used for authenticated calls to /vyprvpn/settings.We found that IP addresses were logged in the DAPI when a user intentionally wrote them into therequest (83868). This is not part of the client software, so it would need to be a leak of information by aprogrammer who decided to use it to their own ends.The retest confirmed that Golden Frog successfully addressed this issue in DAPI.CONFIDENTIAL INFORMATION / LIMIT DISTRIBUTION17
FindingsURL PATH IN DAPI LOGSID83868ComponentDAPISeverityMediumImpact / Skill fffStatusFixedObservationThe application does not take measures to sanitize sensitive information out of the log outputThe request URL is logged as seen in the following request and log:sudo tail -f /var/log/nginx/access.log grep '50\.78'- - - [05/Oct/2018:17:59:23 -0500] "/vyprvpn/connections/50.78.42.169ffff"500 291 "-" "-" "-"curl -i -H 'X-Real-IP: 127.0.0.1' -d '{"disconnect":"0", onnections/50.78.42.169ffff; echoHTTP/1.1 500 INTERNAL SERVER ERRORServer: nginx/1.10.2Date: Fri, 05 Oct 2018 22:59:23 GMTContent-Type: text/htmlContent-Length: 291Connection: close !DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2 Final//EN" title 500 Internal Server Error /title h1 Internal Server Error /h1 p The server encountered an internal error and was unable to complete yourrequest. Either the server is overloaded or there is an error in theapplication. /p While normal requests to /vyprvpn/connections/ ip do not result in a log being created (for anunknown reason), requests that cause an error such as the above are logged in the access log.RecommendationCheck all instance of calling a log writer to make sure that sensitive information such as IP addressesand usernames are not being logged.CONFIDENTIAL INFORMATION / LIMIT DISTRIBUTION18
Control Panel APIControl Panel API is a special server that allows VyprVPN clients to call it to terminate connections, whereever they may be.Threat AnalysisThe Control Panel API node has awareness of a subset of the user activities exposed to the VPN node,notably: Kill switch activationIt is important that VPN not log this user activity as it contains a timestamp and IP address.MethodologyWe logged into the VPN server and reviewed source code for the kill switch and control panel. Wechecked for configuration and logs in the place where nginx is configured to store logs. We additionallylooked in the syslog and other system logs for evidence of users. We looked at processes running toensure that the system did not have any software that would create logs besides nginx, Unicorn, andPython. We checked that the API could not be called directly by users. We made requests to the serverfrom localhost a
disclosure agreement between Leviathan Security and the company identified as “P d F ” No part of this document may be photocopied, reproduced, or translated to another language without the prior written and documented consent of Leviathan Security Group and the d d “P d F ” page. Disclaimer
Application Security Testing (DAST) Origin Analysis / Software Composition Analysis (SCA) Mobile Application Security Testing (MAST) Application Security Testing as a Service (ASTaaS) Correlation Tools Application Security Testing Orchestration (ASTO) Database Security Scanning Test Coverage Analyzers Interactive Application Security Testing .
HPE Secure IoT Application Lifecycle IoT Endpoints Connectivity Edge Computing Visualization IoT Cloud / Platform HPE Security ArcSight (Security Intelligence)HPE Security Fortify (Application Security)HPE Security -Data Security (Voltage/Atalla) HPE Aruba (Communication Security)HPE ADM (Application Delivery Management)HPE ITOM (IT Operations Management)
ISO 27034 . ISO 27001/2: IT Security ISO 27034: Application Security Part 1: Overview & concepts (Nov. 2011) Part 2: Organization normative framework (Aug. 2015) Part 3: Application security management process Part 4: Application security validation Part 5: Protocols and application security controls data structure Part 6: Security guidance for specific .
1 EOC Review Unit EOC Review Unit Table of Contents LEFT RIGHT Table of Contents 1 REVIEW Intro 2 REVIEW Intro 3 REVIEW Success Starters 4 REVIEW Success Starters 5 REVIEW Success Starters 6 REVIEW Outline 7 REVIEW Outline 8 REVIEW Outline 9 Step 3: Vocab 10 Step 4: Branch Breakdown 11 Step 6 Choice 12 Step 5: Checks and Balances 13 Step 8: Vocab 14 Step 7: Constitution 15
New OWASP Top 10 in 2013. Best Quality Application Security OWAPS Top 10 Calculation. Best Quality Application Security OWASP Top 10 Calculation. Best Quality Application Security What works Reallywell? Best Quality Application Security Three Fundamentals to a Security Solution.
Knowledge of OWASP Top 10, threat modelling, SAST and DAST Capturing security requirements of an application in development Defining, maintaining, and enforcing application security best practices Performing manual and automated code review of application Conducting application security testing f
the public–private partnership law review the real estate law review the real estate m&a and private equity review the renewable energy law review the restructuring review the securities litigation review the shareholder rights and activism review the shipping law review the sports law review the tax disputes and litigation review
Adventure Tourism has grown exponentially worldwide over the past years with tourists visiting destinations previously undiscovered. This allows for new destinations to market themselves as truly .
