diff --git a/src/node.cc b/src/node.cc
index 12219b026d63b7..9be6fa5c507196 100644
--- a/src/node.cc
+++ b/src/node.cc
@@ -69,7 +69,6 @@
 #if defined(_MSC_VER)
 #include <direct.h>
 #include <io.h>
-#define strcasecmp _stricmp
 #define getpid GetCurrentProcessId
 #define umask _umask
 typedef int mode_t;
@@ -1350,35 +1349,35 @@ enum encoding ParseEncoding(const char* encoding,
       break;
   }
 
-  if (strcasecmp(encoding, "utf8") == 0) {
+  if (StringEqualNoCase(encoding, "utf8")) {
     return UTF8;
-  } else if (strcasecmp(encoding, "utf-8") == 0) {
+  } else if (StringEqualNoCase(encoding, "utf-8")) {
     return UTF8;
-  } else if (strcasecmp(encoding, "ascii") == 0) {
+  } else if (StringEqualNoCase(encoding, "ascii")) {
     return ASCII;
-  } else if (strcasecmp(encoding, "base64") == 0) {
+  } else if (StringEqualNoCase(encoding, "base64")) {
     return BASE64;
-  } else if (strcasecmp(encoding, "ucs2") == 0) {
+  } else if (StringEqualNoCase(encoding, "ucs2")) {
     return UCS2;
-  } else if (strcasecmp(encoding, "ucs-2") == 0) {
+  } else if (StringEqualNoCase(encoding, "ucs-2")) {
     return UCS2;
-  } else if (strcasecmp(encoding, "utf16le") == 0) {
+  } else if (StringEqualNoCase(encoding, "utf16le")) {
     return UCS2;
-  } else if (strcasecmp(encoding, "utf-16le") == 0) {
+  } else if (StringEqualNoCase(encoding, "utf-16le")) {
     return UCS2;
-  } else if (strcasecmp(encoding, "binary") == 0) {
+  } else if (StringEqualNoCase(encoding, "binary")) {
     return BINARY;
-  } else if (strcasecmp(encoding, "buffer") == 0) {
+  } else if (StringEqualNoCase(encoding, "buffer")) {
     return BUFFER;
-  } else if (strcasecmp(encoding, "hex") == 0) {
+  } else if (StringEqualNoCase(encoding, "hex")) {
     return HEX;
-  } else if (strcasecmp(encoding, "raw") == 0) {
+  } else if (StringEqualNoCase(encoding, "raw") == 0) {
     if (!no_deprecation) {
       fprintf(stderr, "'raw' (array of integers) has been removed. "
                       "Use 'binary'.\n");
     }
     return BINARY;
-  } else if (strcasecmp(encoding, "raws") == 0) {
+  } else if (StringEqualNoCase(encoding, "raws") == 0) {
     if (!no_deprecation) {
       fprintf(stderr, "'raws' encoding has been renamed to 'binary'. "
                       "Please update your code.\n");
diff --git a/src/node_crypto.cc b/src/node_crypto.cc
index 21e4d33cab9316..2d508d2cc3d527 100644
--- a/src/node_crypto.cc
+++ b/src/node_crypto.cc
@@ -24,10 +24,6 @@
 #include <stdlib.h>
 #include <string.h>
 
-#if defined(_MSC_VER)
-#define strcasecmp _stricmp
-#endif
-
 #if OPENSSL_VERSION_NUMBER >= 0x10000000L
 #define OPENSSL_CONST const
 #else
@@ -4232,7 +4228,7 @@ void DiffieHellman::DiffieHellmanGroup(
   for (size_t i = 0; i < arraysize(modp_groups); ++i) {
     const modp_group* it = modp_groups + i;
 
-    if (strcasecmp(*group_name, it->name) != 0)
+    if (!StringEqualNoCase(*group_name, it->name))
       continue;
 
     initialized = diffieHellman->Init(it->prime,
diff --git a/src/util-inl.h b/src/util-inl.h
index 69dc5b2a61a1a9..7051659a5e0e6a 100644
--- a/src/util-inl.h
+++ b/src/util-inl.h
@@ -203,7 +203,19 @@ void SwapBytes(uint16_t* dst, const uint16_t* src, size_t buflen) {
     dst[i] = (src[i] << 8) | (src[i] >> 8);
 }
 
+char ToLower(char c) {
+  return c >= 'A' && c <= 'Z' ? c + ('a' - 'A') : c;
+}
 
+bool StringEqualNoCase(const char* a, const char* b) {
+  do {
+    if (*a == '\0')
+      return *b == '\0';
+    if (*b == '\0')
+      return *a == '\0';
+  } while (ToLower(*a++) == ToLower(*b++));
+  return false;
+}
 
 }  // namespace node
 
diff --git a/src/util.h b/src/util.h
index 7b2bc0f1a270ba..671b98dd25b191 100644
--- a/src/util.h
+++ b/src/util.h
@@ -178,6 +178,12 @@ inline TypeName* Unwrap(v8::Local<v8::Object> object);
 
 inline void SwapBytes(uint16_t* dst, const uint16_t* src, size_t buflen);
 
+// tolower() is locale-sensitive.  Use ToLower() instead.
+inline char ToLower(char c);
+
+// strcasecmp() is locale-sensitive.  Use StringEqualNoCase() instead.
+inline bool StringEqualNoCase(const char* a, const char* b);
+
 class Utf8Value {
   public:
     explicit Utf8Value(v8::Isolate* isolate, v8::Local<v8::Value> value);
diff --git a/test/cctest/util.cc b/test/cctest/util.cc
index fe966d9b34073c..37133aca562b72 100644
--- a/test/cctest/util.cc
+++ b/test/cctest/util.cc
@@ -56,3 +56,21 @@ TEST(UtilTest, ListHead) {
   EXPECT_TRUE(list.IsEmpty());
   EXPECT_FALSE(list.begin() != list.end());
 }
+
+TEST(UtilTest, StringEqualNoCase) {
+  using node::StringEqualNoCase;
+  EXPECT_FALSE(StringEqualNoCase("a", "b"));
+  EXPECT_TRUE(StringEqualNoCase("", ""));
+  EXPECT_TRUE(StringEqualNoCase("equal", "equal"));
+  EXPECT_TRUE(StringEqualNoCase("equal", "EQUAL"));
+  EXPECT_TRUE(StringEqualNoCase("EQUAL", "EQUAL"));
+  EXPECT_FALSE(StringEqualNoCase("equal", "equals"));
+  EXPECT_FALSE(StringEqualNoCase("equals", "equal"));
+}
+
+TEST(UtilTest, ToLower) {
+  using node::ToLower;
+  EXPECT_EQ('0', ToLower('0'));
+  EXPECT_EQ('a', ToLower('a'));
+  EXPECT_EQ('a', ToLower('A'));
+}