components/data_reduction_proxy/browser/data_reduction_proxy_tamper_detection.h

   1 // Copyright 2014 The Chromium Authors. All rights reserved.
   2 // Use of this source code is governed by a BSD-style license that can be
   3 // found in the LICENSE file.
   4
   5 // This file implements the tamper detection logic, which detects whether
   6 // there are middleboxes and whether they are tampering with the response
   7 // which may break correct communication and data transfer between the Chromium
   8 // client and the data reduction proxy.
   9 //
  10 // At a high level, the tamper detection process works in two steps:
  11 // 1. The data reduction proxy selects a fraction of responses to analyze,
  12 //    generates a series of fingerprints for each, and appends them to the
  13 //    Chrome-Proxy response headers;
  14 // 2. The client re-generate the fingerprints using the same method as the
  15 //    proxy, compares them to the fingerprints in the response, and generates
  16 //    UMA. A response is considered to have been tampered with if the
  17 //    fingerprints do not match.
  18 //
  19 // Four fingerprints are generated by the data reduction proxy:
  20 // 1. Fingerprint of the Chrome-Proxy header, which is designed to check
  21 //    whether the Chrome-Proxy header has been modified or not;
  22 // 2. Fingerprint of the Via header, which is designed to check whether there
  23 //    are middleboxes between the Chromium client and the data reduction proxy;
  24 // 3. Fingerprint of a list of headers, which is designed to check whether the
  25 //    values of a list of headers (list is defined by the data reduction proxy)
  26 //    have been modified or deleted;
  27 // 4. Fingerprint of the Content-Length header, which is designed to check
  28 //    whether the response body has been modified or not (the code assumes that
  29 //    different Content-Length values indicate different response bodies).
  30 //
  31 // On the client side, the fingerprint of the Chrome-Proxy header will be
  32 // checked first. If the fingerprint indicates that the Chrome-Proxy header has
  33 // not been modified, then the other fingerprints will be considered to be
  34 // reliable and will be checked next; if not, then it's possible that the other
  35 // fingerprints have been tampered with and thus they will not be checked.
  36 // If middlebox removes all the fingerprints then such tampering will not be
  37 // detected.
  38 //
  39 // Detected tampering information will be reported to UMA. In general, for each
  40 // fingerprint, the client reports the number of responses that have been
  41 // tampered with for different carriers. For the fingerprint of the
  42 // Content-Length header, which indicates whether the response body has been
  43 // modified or not, the reports of tampering are separated by MIME type of the
  44 // response body.
  45
  46 #ifndef COMPONENTS_DATA_REDUCTION_PROXY_BROWSER_DATA_REDUCTION_PROXY_TAMPER_DETECTION_H_
  47 #define COMPONENTS_DATA_REDUCTION_PROXY_BROWSER_DATA_REDUCTION_PROXY_TAMPER_DETECTION_H_
  48
  49 #include <map>
  50 #include <string>
  51 #include <vector>
  52
  53 #include "net/proxy/proxy_service.h"
  54
  55 namespace net {
  56 class HttpResponseHeaders;
  57 }
  58
  59 namespace data_reduction_proxy {
  60
  61 // Detects if the response sent by the data reduction proxy has been modified
  62 // by intermediaries on the Web.
  63 class DataReductionProxyTamperDetection {
  64  public:
  65   // Checks if the response contains tamper detection fingerprints added by the
  66   // data reduction proxy, and determines if the response had been tampered
  67   // with if so. Results are reported to UMA. HTTP and HTTPS traffic are
  68   // reported separately, specified by |scheme_is_https|. Returns true if
  69   // the response has been tampered with.
  70   static bool DetectAndReport(const net::HttpResponseHeaders* headers,
  71                               bool scheme_is_https);
  72
  73   // Tamper detection checks |response_headers|. Histogram events are reported
  74   // by |carrier_id|; |scheme_is_https| determines which histogram to report
  75   // (HTTP and HTTPS are reported separately).
  76   DataReductionProxyTamperDetection(
  77       const net::HttpResponseHeaders* response_headers,
  78       bool scheme_is_https,
  79       unsigned carrier_id);
  80
  81   virtual ~DataReductionProxyTamperDetection();
  82
  83  private:
  84   FRIEND_TEST_ALL_PREFIXES(DataReductionProxyTamperDetectionTest,
  85                            TestFingerprintCommon);
  86   FRIEND_TEST_ALL_PREFIXES(DataReductionProxyTamperDetectionTest,
  87                            ChromeProxy);
  88   FRIEND_TEST_ALL_PREFIXES(DataReductionProxyTamperDetectionTest,
  89                            Via);
  90   FRIEND_TEST_ALL_PREFIXES(DataReductionProxyTamperDetectionTest,
  91                            OtherHeaders);
  92   FRIEND_TEST_ALL_PREFIXES(DataReductionProxyTamperDetectionTest,
  93                            ContentLength);
  94   FRIEND_TEST_ALL_PREFIXES(DataReductionProxyTamperDetectionTest,
  95                            HeaderRemoving);
  96   FRIEND_TEST_ALL_PREFIXES(DataReductionProxyTamperDetectionTest,
  97                            ValuesToSortedString);
  98   FRIEND_TEST_ALL_PREFIXES(DataReductionProxyTamperDetectionTest,
  99                            GetHeaderValues);
 100   FRIEND_TEST_ALL_PREFIXES(DataReductionProxyTamperDetectionTest,
 101                            DetectAndReport);
 102
 103   // Returns the result of validating Chrome-Proxy header.
 104   bool ValidateChromeProxyHeader(const std::string& fingerprint) const;
 105
 106   // Reports UMA for tampering of the Chrome-Proxy header.
 107   void ReportUMAforChromeProxyHeaderValidation() const;
 108
 109   // Returns the result of validating the Via header. |has_chrome_proxy|
 110   // indicates that the data reduction proxy's Via header occurs or not.
 111   bool ValidateViaHeader(const std::string& fingerprint,
 112                          bool* has_chrome_proxy_via_header) const;
 113
 114   // Reports UMA for tampering of the Via header.
 115   void ReportUMAforViaHeaderValidation(bool has_chrome_proxy_via_header) const;
 116
 117   // Returns the result of validating a list of headers.
 118   bool ValidateOtherHeaders(const std::string& fingerprint) const;
 119
 120   // Reports UMA for tampering of values of the list of headers.
 121   void ReportUMAforOtherHeadersValidation() const;
 122
 123   // Returns the result of validating the Content-Length header.
 124   bool ValidateContentLengthHeader(const std::string& fingerprint) const;
 125
 126   // Reports UMA for tampering of the Content-Length header.
 127   void ReportUMAforContentLengthHeaderValidation() const;
 128
 129   // Returns a string representation of |values|.
 130   static std::string ValuesToSortedString(std::vector<std::string>* values);
 131
 132   // Returns raw MD5 hash value for a given string |input|. It is different to
 133   // base::MD5String which is base16 encoded.
 134   static void GetMD5(const std::string& input, std::string* output);
 135
 136   // Returns all the values of |header_name| of the response |headers| as a
 137   // vector. This function is used for values that need to be sorted later.
 138   static std::vector<std::string> GetHeaderValues(
 139       const net::HttpResponseHeaders* headers,
 140       const std::string& header_name);
 141
 142   // Pointer to response headers.
 143   const net::HttpResponseHeaders* response_headers_;
 144
 145   // If true, the connection to the data reduction proxy is over HTTPS;
 146   const bool scheme_is_https_;
 147
 148   // Carrier ID: the numeric name of the current registered operator.
 149   const unsigned carrier_id_;
 150
 151   DISALLOW_COPY_AND_ASSIGN(DataReductionProxyTamperDetection);
 152 };
 153
 154 }  // namespace data_reduction_proxy
 155 #endif  // COMPONENTS_DATA_REDUCTION_PROXY_BROWSER_DATA_REDUCTION_PROXY_TAMPER_DETECTION_H_