Web Proxy Auto Discovery (WPAD)

Back in the mid-aughts, Adam G., a colleague on the IE team, used the email signature “IE Networking Team – Without us, you’d be browsing your hard drive.” And while I’m sure it was meant to be a bit tongue-in-cheek, it’s really true– without a working network stack, web browsers aren’t nearly as useful.

Background on Proxy Determination

One of the very first things a browser must do on startup is figure out how to send requests over the network. Typically, the host operating system already provides the transport (TCP/IP, UDP) and lower-level primitives, so the browser’s first task is to figure out whether or not web requests should be sent through a proxy. Until this question is resolved, the browser cannot send any network requests to load pages, sync profile information, update phishing blocklists, etc.

In some cases, proxy determination is simple— the browser is directly configured to ignore proxies, or to send all requests to a directly specified proxy.

However, for convenience and to simplify cases where a user might move a laptop between different networks with different proxy requirements, all major browsers support an algorithm called “Web Proxy Auto Discovery”, or WPAD. The WPAD process is meant to find and download a Proxy AutoConfiguration Script (PAC) for the current network.

The steps of the WPAD protocol are straightforward, if lengthy:

  1. Determine whether WPAD should be used, either by looking at browser settings or asking the host operating system if the browser is configured to match the OS setting.
  2. Ensure the network is ready.
  3. If WPAD is to be used, issue a DHCPINFORM query to ask for the URL of the PAC script to use.
  4. If the DHCPINFORM query fails to return a URL, perform a DNS lookup for the unqualified hostname wpad.
  5. If the DNS lookup succeeds, then the PAC URL shall be http://wpad/wpad.dat.
  6. Establish a HTTP(S) connection to discovered URL’s server and download the PAC script.
  7. If the PAC script downloads successfully, parse and optionally compile it.
  8. For each network request, call FindProxyForURL() in the PAC script and use the proxy settings returned from the function.

While conceptually simple, any of these steps might fail, and any failure might prevent the browser from using the network.


… or “Why on earth do I see Downloading proxy script… for a few seconds every time I start my browser!??!”

A Microsoft Edge feature team reached out to the networking team this week asking for help with an observed 3 second delay in the initialization of their feature. They observed that this delay magically disappeared if Fiddler happened to be running.

With symptoms like that, proxy determination is the obvious suspect, because Fiddler specifies the exact proxy configuration for browsers to use, meaning that they do not need to perform the WPAD process.

We asked the team to take an Edge network trace using the “Capture on Startup” steps. Sure enough, when we analyzed the resulting NetLog, we found almost exactly three seconds of blocking time during startup:

             --> new_config = Auto-detect
              --> source = "WPAD DHCP"
            --> net_error = -348 (ERR_PAC_NOT_IN_DHCP) 
              --> host = "wpad:80" 

Note: Timestamps [e.g. t=52] are shown in milliseconds.

Because the browser took a full three seconds to decide whether or not to use a proxy, every feature that relies on the network will take at least three seconds to get the data it needs.

So, where’s the delay coming from? In this case, the delay comes from two places: a two second delay for PAC_FILE_DECIDER_WAIT and a one second delay for the DNS lookup of wpad.

The two second PAC_FILE_DECIDER_WAIT [Step #2] is a deliberate delay that is meant to delay PAC lookups after a network change event is observed, to accommodate situations where the browser is notified of a network change by the Operating System before the network is truly “ready” to perform the DHCP/DNS/Download steps of WPAD. In this browser-startup case, we haven’t yet figured out why the browser thinks a network change has occurred, but the repro is not consistent and it seems likely to be a bug.

The (failing) DNS lookup [Step #3] might’ve taken even longer to return, but it timed out after one second thanks to an enabled-by-default feature called WPADQuickCheckEnabled.

This three second delay on startup is bad, but it could be even worse. We got reports from one Microsoft employee that every browser startup took around 21 seconds to navigate anywhere. In looking at his network log, we found that the wpad DNS lookup [Step #5] succeeded, returning an IP address, but the returned IP was unreachable and took 21 seconds to timeout during TCP/IP connection establishment.

What makes these delays especially galling is that they were all encountered on a network that does not actually need a proxy!


Beyond the time delays, each of these steps might fail, and if a proxy is required on the current network, the user will be unable to browse until the problem is corrected.

For example, we recently saw that [Step #7] failed for some users because the Utility Process running the PAC script always crashed due to forbidden 3rd-party code injection. When the Utility Process crashes, Chromium attempts to bypass the proxy and send requests directly to the server, which was forbidden by the Enterprise customer’s network firewall.

We’ve also found that care must be taken in the JavaScript implementation of FindProxyForURL() [Step #8] because script functions behave slightly differently across different browsers. In most cases, scripts work just fine across browsers, but sometimes corner cases are encountered that require careful handling.

Script Download

In Chromium, if a PAC script must be downloaded, it is fetched bypassing the cache.

Even if we were to comment out the LOAD_DISABLE_CACHE directive in the fetch, this wouldn’t allow reuse of a previously downloaded script file– my assumption is that the download is happening in a NetworkContext that doesn’t actually have a persistent cache, but I haven’t looked into this.

The PAC script fetches will be repeated on network change or browser restart.


WPAD is something of a security threat, because it means that another computer on your network might be able to become your proxy server without you realizing it. In theory, HTTPS traffic is protected against malicious proxy servers, but non-secure HTTP traffic hasn’t yet been eradicated from the web, and users might not notice if a malicious proxy performed an SSLStripping attack on a site that wasn’t HSTS preloaded, for example.

Note: Back in 2016, it was noticed that the default Chromium proxy script implementation leaked full URLs (including HTTPS URLs’ query strings) to the proxy script; this was fixed by truncating the URL to the hostname. (In the new world of DoH, there’s some question as to whether we might be able to avoid sending the hostname to the proxy at all).

Edge Legacy and Internet Explorer have a surprising default behavior that treats sites for which a PAC script returns DIRECT (“bypass the proxy for this request“) as belonging to your browser’s Intranet Zone.

This mapping can lead to functionality glitches and security/privacy risks. Even in Chrome and the new Edge, Windows Integrated Authentication still occurs Automatically for the Windows Intranet Zone, which means this WPAD Zone Mapping behavior is still relevant in modern browsers.

Chrome performing Automatic Authentication due to Proxy Bypass

Edge Legacy and Internet Explorer

Interestingly, performance and functionality problems with WPAD might have been less common for the Edge Legacy and Internet Explorer browsers on Windows 10. That’s because both of these browsers rely upon the WinHTTP Web Proxy Auto-Discovery Service:

This is a system service that handles proxy determination tasks for clients using the WinHTTP/WinINET HTTP(S) network stacks. Because the service is long-running, performance penalties are amortized (e.g. a 3 second delay once per boot is much cheaper than a 3 second delay every time your browser starts), and the service can maintain caches across different processes.

Chromium does not, by default, directly use this service, but it can be directed to do so by starting it with the command-line argument:


A Group Policy that matches the command-line argument is also available.


Prior to the enhancement of the WinHTTP WPAD Service, a feature called SmartWPAD was introduced in Internet Explorer 8’s version of WinINET. SmartWPAD caches in the registry a list of networks on which WPAD has not resulted in a PAC URL, saving clients the performance cost of performing the WPAD process each time they restarted for the common case where WPAD fails to discover a PAC file:

Cache entries would be maintained for a given network fingerprint for one month. Notably, the SmartWPAD cache was only updated by WinINET, meaning you’d only benefit if you launched a WinINET-based application (e.g. IE) at least once a month.

When a client (including IE, Chrome, Microsoft Edge, Office, etc) subsequently asks for the system proxy settings, SmartWPAD checks if it had previously cached that WPAD was not available on the current network. If so, the API “lies” and says that the user has WPAD disabled.

The SmartWPAD feature still works with browsers running on Windows 7 today.

Notably, it does not seem to function in Windows 10; the registry cache is empty. My Windows 10 Chromium browsers spend ~230ms on the WPAD process each time they are fully restarted.

Update: The WinINET team confirms that SmartWPAD support was removed after Windows 7; for clients using WinINET/WinHTTP it wasn’t needed because they were using the proxy service. Clients like Chromium and Firefox that query WinINET for proxy settings but use their own proxy resolution logic will need to implement a SmartWPAD-like feature optimize performance.

Disabling WPAD

If your computer is on a network that doesn’t need a proxy, you can ensure maximum performance by simply disabling WPAD in the OS settings.

By default (if not overridden by policy or the command line), Chromium adopts the Windows proxy settings by calling WinHttpGetIEProxyConfigForCurrentUser.

On Windows, you can thus turn off WPAD by default by using the Internet Control Panel (inetcpl.cpl) Connections > LAN Settings dialog, or the newer Windows 10 Settings applet’s Automatic Proxy Setup section:

Simply untick the box and browsers that inherit their default settings from Windows (Chrome, Microsoft Edge, Edge Legacy, Internet Explorer, and Firefox) will stop trying to use WPAD.

Looking forward

WPAD is convenient, but somewhat expensive for performance and a bit risky for security/privacy. Every few years, there’s a discussion about disabling it by default (either for everyone, or for non-managed machines), but thus far none of those conversations has gone very far.

Ultimately, we end up with an ugly tradeoff– no one wants to land a change that results in users being limited to browsing their hard drives.

If you’re an end user, consider unticking the “Automatically Detect Settings” checkbox in your Internet settings. If you’re an enterprise administrator, consider deploying a policy to disable WPAD for your desktop fleet.


Avoiding Unexpected Navigation

For over twenty years, browsers broadly supported two features that were often convenient but sometimes accidentally invoked, leading to data loss.

The first feature was that hitting backspace would send the user back one page in their navigation history. This was convenient for those of us who keep our hands on the keyboard, but could lead to data loss– for instance, when filling out a web form, if focus accidentally left a text box, hitting backspace could result in navigating away from the form. Smart websites would warn you with an OnBeforeUnload handler, and some browsers tried to restore the contents of the form if the user understood what happened and hit “forward”, but these protections are imperfect and might not work for all forms.

One of the IE browser UI leads complained about this behavior annually for a decade, and users periodically howled as they lost work. Finally, circa 2016, this feature was removed from Chrome and Microsoft Edge followed in 2018. If you happened to love the old behavior and accept the risk of data loss, you can restore it via extension or in Edge 86, via the edge://flags/#edge-backspace-key-navigate-page-back flag.

The second feature was drop to navigate. Dragging and dropping a file into the browser’s content area (the body of the page) would, (unless the page’s JavaScript was designed to handle the drop, e.g. to upload it or process it locally), navigate to that local file in the current tab. Some folks like that behavior– e.g. web developers loading HTML files from their local filesystem, but it risks the same data loss problem. If a web page doesn’t accept file uploads via drag/drop, the contents of that page will be blown away by navigation. Back in 2015, a bug was filed against Chromium suggesting that the default behavior was too dangerous, and many examples were provided where the default behavior could be problematic. Yesterday, I landed a tiny change for Chromium 85 [later merged to v84] such that dropping a file or URL into the content area of a tab will instead open the file in a new tab:

Dropping in the content area now opens it in a new tab:

If you do want to replace the content of the tab with the dropped file, you can simply drag/drop the file to the tab strip.

A small white arrow shows you which tab’s content will be replaced:

Dropping the file between tabs on the tab strip will insert a new tab at the selected location:

Chrome (85.0.4163/v84) and Microsoft Edge (85.0.541) include this change; it was also later merged to v84. Microsoft Edge Legacy didn’t support drop to navigate. Firefox still has the old behavior; the closest bug I could find is this one. Safari 13.1.1 still has the old behavior and replaces the content of the current tab with the local file. Safari Tech Preview 13.2 Release 108 instead navigates the tab to an error page (NSPOSIXErrorDomain:1 Operation not permitted”).


Browser Basics: User Gestures

The Web Platform offers a great deal of power, and unfortunately evil websites go to great lengths to abuse it. One of the weakest (but simplest to implement) protections against such abuse is to block actions that were not preceded by a “User Gesture.” Such gestures (sometimes more precisely called User Activations) include a variety of simple actions, from clicking the mouse to typing a key; each interpreted as “The user tried to do something in this web content.”

A single user gesture can unlock any of a surprisingly wide array of privileged (“gated”) actions:

  • Allow a popup window to open
  • Allow an Application Protocol to be invoked
  • Allow an OnBeforeUnload dialog box to show
  • Allow the Vibration API to vibrate the device
  • Allow script to take the window fullscreen
  • Allow the password manager to fill the username/password into the page in a way that JavaScript can see them
  • Allow the page to prompt the user for a file to upload
  • Impact the behavior of file downloads (e.g. prompting)
  • and many more

So, when you see a site show a UI like this:

…chances are good that what they’re really trying to do is trick you into performing a gesture (mouse click) so they can perform a privileged action– in this case, open a popup ad in a new tab.

In terms of which actions can cause a gesture, the list is surprisingly limited, and includes mousedown (but not mouseup/click):

// Returns |true| if |type| is the kind of user input that should trigger user interaction observers.
bool IsUserInteractionInputType(blink::WebInputEvent::Typetype) {
// Ideally, this list would be based more off of
// https://whatwg.org/C/interaction.html#triggered-by-user-activation. return type ==
blink::WebInputEvent::Type::kMouseDown ||
type == blink::WebInputEvent::Type::kGestureScrollBegin ||
type == blink::WebInputEvent::Type::kTouchStart ||
type == blink::WebInputEvent::Type::kRawKeyDown; }

Some gestures are considered “consumable”, meaning that a single user action allows only one privileged action; subsequent privileged actions require another gesture. Web Developers do not have unlimited time to consume the action: In Chrome, when you click in a web page, the browser considers this “User Activation” valid for five seconds (as of Februrary 2019) before it expires. Here’s a simple test.

Unfortunately, even this weak protection is subject to both false positives (an unwanted granting of privilege) and false negatives (an action is unexpectedly blocked).

You can learn more about this topic (and the complexity of dealing with nested frames, etc) in the original Chromium User Activation v2 spec, and the User-Activation section of HTML5.


A bit of GREASE keeps the web moving

For the first few years of the web, developers pretty much coded whatever they thought was cool and shipped it. Specifications, if written at all, were an afterthought.

Then, for the next two decades, spec authors drafted increasingly elaborate specifications with optional features and extensibility points meant to be used to enable future work.

Unfortunately, browser and server developers often only implemented enough of the specs to ensure interoperability, and rarely tested that their code worked properly in the face of features and data allowed by the specs but not implemented in the popular clients.

Over the years, the web builders started to notice that specs’ extensibility points were rusting shut– if a new or upgraded client tried to make use of a new feature, or otherwise change what it sent as allowed by the specs, existing servers would fail when they saw the encountered the new values.

In light of this, spec authors came up with a clever idea: clients should send random dummy values allowed by the spec, causing spec-non-compliant servers that fail to properly ignore those values to fail immediately. This concept is called GREASE (with the backronym “Generate Random Extensions And Sustain Extensibility“), and was first implemented for the values sent by the TLS handshake. When connecting to servers, clients would claim to support new ciphersuites and handshake extensions, and intolerant servers would fail. Users would holler, and engineers could follow up with the broken site’s authors and developers about how to fix their code. To avoid “breaking the web” too broadly, GREASE is typically enabled experimentally at first, in Canary and Dev channels. Only after the scope of the breakages is better understood does the change get enabled for most users.

GREASE has proven such a success for TLS handshakes that the idea has started to appear in new places. Last week, the Chromium project turned on GREASE for HTTP2 in Canary/Dev for 50% of users, causing connection failures to many popular sites, including some run by Microsoft. These sites will need to be fixed in order to properly load in the new builds of Chromium.

// Enable "greasing" HTTP/2, that is, sending SETTINGS parameters with reserved identifiers and sending frames of reserved types, respectively. If greasing Frame types, an HTTP/2 frame with a reserved frame type will be sent after every HEADERS and SETTINGS frame. The same frame will be sent out on all connections to prevent the retry logic from hiding broken servers.
NETWORK_SWITCH(kHttp2GreaseSettings, "http2-grease-settings") NETWORK_SWITCH(kHttp2GreaseFrameType, "http2-grease-frame-type")

One interesting consequence of sending GREASE H2 Frames is that it requires moving the END_STREAM flag (recorded as fin=true in the netlog) from the HTTP2_SESSION_SEND_HEADERS frame into an empty (size=0) HTTP2_SESSION_SEND_DATA frame; unfortunately, the intervening GREASE Frame is not presently recorded in the netlog.

You can try H2 GREASE in Chrome Stable using command line flags that enable GREASE settings values and GREASE frames respectively:

chrome.exe --http2-grease-settings bing.com
chrome.exe --http2-grease-frame-type bing.com

Alternatively, you can disable the experiment in Dev/Canary:

chrome.exe --disable-features=Http2Grease

GREASE is baked into the new HTTP3 protocol (Cloudflare does it by default) and the new UA Client Hints specification (where it’s blowing up a fair number of sites). I expect we’ll see GREASE-like mechanisms appearing in most new web specs where there are optional or extensible features.



Someone complained that a Japanese page is garbled in Edge/Chrome, but renders with the correct characters in Firefox and IE:

The problem is that Chromium is using an unexpected character set to interpret the response in the HTML Parser. That happens because the server doesn’t send a proper character set directive. To avoid problems like this and improve performance, document authors should specify the character set in the HTTP response headers:

Content-Type: text/html; charset=utf-8

If a charset isn’t specified in the headers, Chrome looks for a META CHARSET declaration within the response body text (“note the absurdity of encoding the character encoding in the document that you’re trying to decode“). The HTML5 spec demands that documents be encoded in UTF-8, and that the charset declaration, if any, appears within the first 1024 bytes of the response.

Chrome checks the full head for a character-set directive, and if it doesn’t find one, ensures that it’s looked through at least 1024 bytes before giving up.

Unfortunately, this site accidentally includes a div tag up in the head (ending the head section prematurely), and buries the charset down 1479 bytes into the response:

To avoid problems like this:

  1. Specify the CHARSET in the Content-Type response header, and
  2. Ensure META CHARSET appears as the first element of your HEAD.
  3. To avoid problems in legacy browsers, write it as utf-8 rather than as utf8.

Get your <HEAD> in order!


Client Certificate Authentication

While most HTTPS sites only authenticate the server (using a certificate sent by the website), HTTPS also supports a mutual authentication mode, whereby the client supplies a certificate that authenticates the visiting user’s identity. Such a certificate might be stored on a SmartCard, or used as a part of an OS identity feature like Windows Hello.

To request mutual authentication, servers send a CertificateRequest message to the client during the HTTPS handshake, specifying a criteria filter that the browser will use to find a client certificate to satisfy the server’s request.

If a client certificate is supplied in the browser’s Certificate response to the server’s challenge, the browser proves the user’s possession of that certificate using the private key that matches that client certificate’s public key.

A client may choose not to send a certificate (either because no matching certificate is available, or because the user declined to supply a certificate that it had)—in such cases, the server may terminate the handshake (showing a Client Certificate Required error message) or it may continue the handshake and attempt to authenticate the user via other means.

Certificate Selection

The CertificateRequest message allows the server to specify criteria for the certificates it is willing to accept from the client, including details such as the certificate’s issuer, and key/signature/hash types.

The browser consults the Operating System’s trust store (Keychain on Mac OS X, certmgr.msc on Windows) to find any candidate certificates (unexpired certificates with the Client Authentication purpose set and a private key available) that match the server-supplied filtering criteria:

The private key for a given certificate might be stored on a SmartCard — when a SmartCard is inserted, the certificate(s) on it are “virtually” propagated to the OS trust store for use by browsers and other applications.

Certificates that meet the server’s filtering criteria are shown in a prompt:

If the user hits “Cancel”, the handshake is completed without sending a certificate. However, if the user selects a certificate, the browser caches that decision for the lifetime of the browser instance. The selected certificate will be resent on all new connections to the target origin and the prompt will not be shown again.

Today, there’s no good way to clear the selection decision, short of restarting the browser entirely. In contrast, legacy IE offered two very awkward mechanisms, the Clear SSLState button in the Internet Control Panel, and the ClearAuthenticationCache web API.

Automatic Selection of Client Certificate

Internet Explorer and Edge Legacy offered a behavior (Don’t prompt for client certificate selection when only one certificate exists, URLACTION_CLIENT_CERT_PROMPT), on-by-default for the Local Intranet Zone:

…whereby the browser would not prompt the user to select a certificate if the user only has one certificate that matches the server’s request. In such cases, the client would automatically send the matching certificate without showing a prompt.

For other zones, IE and Edge Legacy do prompt the user to select a certificate before any certificate is sent. This is a privacy measure, because if the browser silently sends the user’s identity to any website that asks for it, this is a “super-cookie” that would allow tracking that user across sites. Also, the client’s certificate might directly contain personally identifiable information about the user (e.g. their email address, office phone number, home address, etc).

Chromium (and thus Chrome, Edge, Brave, Opera, Vivaldi) largely does not use the concept of Zones, so instead the AutoSelectCertificateForUrls policy exists. This policy allows an IT administrator to configure clients to automatically send certificates to specified websites that request them, which can be used to satisfy the need to have, say, the user’s Windows Hello certificate sent to *.login.microsoft.com sites.

Here are two examples: the first selects the first certificate issued by “Windows Hello PIN – MSIT1” and the second rule selects the certificate with a SubjectCN=”RSACSP”.

If you’re trying to set a rule whereby multiple client certificates are valid candidates and the client should just return the first found match, just add another rule with the same pattern and a different filter.

For instance, this set will use SubjectCN=”RSACP” if a matching certificate found, or a certificate with IssuerCN=”Windows Hello PIN – MSIT1” if not:

A screenshot of a cell phone

Description automatically generated

However, as you may have noticed, the AutoSelectCertificateForUrls policy has one significant limitation, which is that it always sends the user’s first matching certificate to the selected site. Some users might have more than one certificate that matches the policy (for instance, some enterprises have both “test” and “production” certificates.

To address this shortcoming, the Edge team introduced a new policy in Edge 81. The new ForceCertificatePromptsOnMultipleMatches policy which does as it says: If the client has multiple certificates that could be used to satisfy the {OriginFilter->CertificateFilter} policy specified by a AutoSelectCertificateForUrls policy, instead of simply sending the first matching certificate, the browser will instead show a certificate selection prompt filtered to the certificates that match the policy.

If you find that Microsoft Edge shows a client certificate selection prompt in one scenario where other browsers do not, one possibility is that the site in question is not actually requesting a client certificate from those other browsers for some reason. For instance, some web authentication flows, including Microsoft’s AAD login, take the browser’s User-Agent into account when deciding what authentication mechanisms to use with the client.

In order to understand exactly what’s going on with Client Authentication, collect Network Traffic logs. SSL_HANDSHAKE_MESSAGE_RECEIVED messages of type 13 represent the client certificate request.


Bonus trivia

  1. Some smart people persuasively argue that mutually-authenticated TLS is a misfeature.
  2. Registry-specified certificate selection policies apply across browser profiles, meaning that they are in force even when the user is in an Incognito browser session (!!)
  3. A user’s certificate selection apply for the entire browser session, making logout for ClientCert-authenticated sites an unsolved problem.
  4. Client Certificate prompting behavior on Android is weird.
  5. Client Certificate authentication is generally not available in 3rd-party browsers on iOS (Safari has access to the keychain).
  6. There are some test servers available.

The general notion of “how Client Certificates were supposed to work” was that each user would have one certificate for each organization to which they belong, issued by that organization’s root certificate. When visiting that organization’s servers, the server would send in the CertificateRequest message the identifier(s) of the root certificate(s) to which acceptable client certificates chain (using the certificate_authorities structure). The visiting client would then filter the certificates available for selection to only those that chain to that root (hopefully one certificate).

So, say I have two certificates, e.g. USA-NationalID and Microsoft-EmployeeID. When I visit https://portal.microsoft.com, Microsoft sends a CertificateRequest with a MicrosoftRootCA in the certificate_authorities field. My browser automatically filters my client certificates list to just the Microsoft-EmployeeID certificate and then sends that. In contrast, when I visit https://irs.gov, the government sends a CertificateRequest with a USGovernmentRootCA in the certificate_authorities field. My browser automatically filters my client certificates list to just the USA-NationalID certificate and sends that.

In practice, unfortunately, things haven’t worked out that way. Most organizations have not had the infrastructure or discipline to configure things to work like that, and as a consequence you end up with varying client behavior.

Firefox doesn’t seem to filter the certificate list, but it does offer a “Remember this decision” checkbox which presumably reduces user annoyance:

Firefox does not respect the Windows Trust Store, so each client certificate must be manually loaded into Firefox’s configuration. This is a hassle, but it tends to result in a somewhat “cleaner” experience where the user isn’t distracted by random certificates that might be cluttering Windows’ cert store.

In some cases, organizations are generating invalid client certificates but expecting them to work, leading us to create compat accommodations like the FEATURE_CLIENTAUTHCERTFILTER Feature Control Key.

In the browser, SmartCards can be used for two ways: HTTPS Client Certificate Authentication, and Windows Integrated Authentication.

  • Straight TLS mutual authentication, as described above.
  • Windows Integrated Authentication that occurs when visiting a website that sends a WWW-Authenticate: Negotiate header. The client may automatically send the user’s login credentials (Intranet Zone). Or, if those creds do not work or the Zone is not configured for automatic credential release (Non-intranet), the user will be prompted for credentials to use. In Edge 79, the user would get a prompt with two blank fields (“Username” and “Password”). In Edge 80 or later, upon noticing that the user has configured Windows Hello, the user will be shown the Windows Hello auth dialog that allows the user to use their face, type a PIN, use a SmartCard, etc. So, now Edge 80 matches Edge Legacy (v18 and lower).

Low Level Details 1
Low Level Details 2

Nice discussion (with pictures) of setting up client cert auth on IIS.

In Windows 10 Apps, the AppContainer must have the sharedUserCertificates capability to use certificates from the trust store.

This is the first message the client can send after receiving a ServerHelloDone message. This message is only sent if the server requests a certificate. If no suitable certificate is available, the client MUST send a certificate message containing no certificates. That is, the certificate_list structure has a length of zero. If the client does not send any certificates, the server MAY at its discretion either continue the handshake without client authentication, or respond with a fatal handshake_failure alert. Also, if some aspect of the certificate chain was unacceptable (e.g., it was not signed by a known, trusted CA), the server MAY at its discretion either continue the handshake (considering the client unauthenticated) or send a fatal alert.

CertificateVerify signs using the client certificate’s private key.

CertOpenStore “my” store


ClientAuthIssuer trust store.

Hard Problems: Fetch in Serviceworker scenario — how can the user select a certificate when no UI is allowed?

Analyzing Network Traffic Logs (NetLog json)

Previously, I’ve described how to capture a network traffic log from Microsoft Edge, Google Chrome, and applications based on Chromium or Electron.

In this post, I aim to catalog some guidance for looking at these logs to help find the root cause of captured problems and otherwise make sense of the data collected.

Last Update: April 24, 2020I expect to update this post over time as I continue to gain experience in analyzing network logs.

Choose A Viewer – Fiddler or Catapult

After you’ve collected the net-export-log.json file using the about:net-export page in the browser, you’ll need to decide how to analyze it.

The NetLog file format consists of a JSON-encoded stream of event objects that are logged as interesting things happen in the network layer. At the start of the file there are dictionaries mapping integer IDs to symbolic constants, followed by event objects that make use of those IDs. As a consequence, it’s very rare that a human will be able to read anything interesting from a NetLog.json file using just a plaintext editor or even a JSON parser.

The most common (by far) approach to reading NetLogs is to use the Catapult NetLog Viewer, a HTML/JavaScript application which loads the JSON file and parses it into a much more readable set of events.

An alternative approach is to use the NetLog Importer for Telerik Fiddler.

Importing NetLogs to Fiddler

For Windows users who are familiar with Fiddler, the NetLog Importer extension for Fiddler is easy-to-use and it enables you to quickly visualize HTTP/HTTPS requests and responses. The steps are easy:

  1. Install the NetLog Importer,
  2. Open Fiddler, ideally in Viewer mode fiddler.exe -viewer
  3. Click File > Import > NetLog JSON
  4. Select the JSON file to import

In seconds, all of the HTTP/HTTPS traffic found in the capture will be presented for your review. If the log was compressed before it was sent to you, the importer will automatically extract the first JSON file from a chosen .ZIP or .GZ file, saving you a step.

In addition to the requests and responses parsed from the log, there are a number of pseudo-Sessions with a fake host of NETLOG that represent metadata extracted from the log:

These pseudo-sessions include:

  • RAW_JSON contains the raw constants and event data. You probably will never want to examine this view.
  • CAPTURE_INFO contains basic data about the date/time of the capture, what browser and OS version were used, and the command line arguments to the browser.
  • ENABLED_EXTENSIONS contains the list of extensions that are enabled in this browser instance. This entry will be missing if the log was captured using the –log-net-log command line argument.
  • URL_REQUESTS contains a dictionary mapping every event related to URL_REQUEST back to the URL Requests to which it belongs. This provides a different view of the events that were used in the parsing of the Web Sessions added to the traffic list.
  • SECURE_SOCKETS contains a list of all of the HTTPS sockets that were established for network requests, including the certificates sent by the server and the parameters requested of any client certificates. The Server certificates can be viewed by saving the contents of a –BEGIN CERTIFICATE– entry to a file named something.cer. Alternatively, select the line, hit CTRL+C, click Edit > Paste As Sessions, select the Certificates Inspector and press its Content Certificates button.

You can then use Fiddler’s UI to examine each of the Web Sessions.


The NetLog format currently does not store request body bytes, so those will always be missing (e.g. on POST requests).

Unless the Include Raw Bytes option was selected by the user collecting the capture, all of the response bytes will be missing as well. Fiddler will show a “dropped” notice when the body bytes are missing:

If the user did not select the Include Cookies and Credentials option, any Cookie or Authorization headers will be stripped down to help protect private data:

Scenario: Finding URLs

You can use Fiddler’s full text search feature to look for URLs of interest if the traffic capture includes raw bytes. Otherwise, you can search the Request URLs and headers alone.

On any session, you can use Fiddler’s “P” keystroke (or the Select > Parent Request context menu command) to attempt to walk back to the request’s creator (e.g. referring HTML page).

You can look for the traffic_annotation value that reflects why a resource was requested by looking for the X-Netlog-Traffic_Annotation Session Flag.

Scenario: Cookie Issues

If Fiddler sees that cookies were not set or sent due to features like SameSiteByDefault cookies, it will make a note of that in the Session using a psuedo $NETLOG-CookieNotSent or $NETLOG-CookieNotSet header on the request or response:

Closing Notes

If you’re interested in learning more about this extension, see announcement blog post and the open-source code.

While the Fiddler Importer is very convenient for analyzing many types of problems, for others, you need to go deeper and look at the raw events in the log using the Catapult Viewer.

Viewing NetLogs with the Catapult Viewer

Opening NetLogs with the Catapult NetLog Viewer is even simpler:

  1. Navigate to the web viewer
  2. Select the JSON file to view

If you find yourself opening NetLogs routinely, you might consider using a shortcut to launch the Viewer in an “App Mode” browser instance: msedge.exe --app=https://netlog-viewer.appspot.com/#import

The App Mode instance is a standalone window which doesn’t contain tabs or other UI:

Note that the Catapult Viewer is a standalone HTML application. If you like, you can save it as a .HTML file on your local computer and use it even when completely disconnected from the Internet. The only advantage to loading it from appspot.com is that the version hosted there is updated from time to time.

Along the left side of the window are tabs that offer different views of the data– most of the action takes place on the Events tab.


If the problem only exists on one browser instance, check the Command Line parameters and Active Field trials sections on the Import tab to see if there’s an experimental flag that may have caused the breakage. Similarly, check the Modules tab to see if there are any browser extensions that might explain the problem.

Each URL Request has a traffic_annotation value which is a hash you can look up in annotations.xml. That annotation will help you find what part of Chromium generated the network request:

Most requests generated by web content will be generated by the blink_resource_loader, navigations will have navigation_url_loader and requests from features running in the browser process are likely to have other sources.

Scenario: DNS Issues

Look at the DNS tab on the left, and HOST_RESOLVER_IMPL_JOB entries in the Events tab.

One interesting fact: The DNS Error page performs an asynchronous probe to see whether the configured DNS provider is working generally. The Error page also has automatic retry logic; you’ll see a duplicate URL_REQUEST sent shortly after the failed one, with the VALIDATE_CACHE load flag added to it. In this way, you might see a DNS_PROBE_FINISHED_NXDOMAIN error magically disappear if the user’s router’s DNS flakes.

You may see a UDP_SOCKET for the address fd3e:4f5a:5b81::1 and wonder why there is a connection to a server with the hostname dns.msftncsi.com. This is a red-herring; this UDP socket is meant to be used only to determine information about the local network configuration (to find out the default local address on a multi-homed endpoint). No packets should be sent for this socket. Upstream in Chrome/Chromium, Google uses an address it controls for the same purpose.

Scenario: Cookie Issues

Look for COOKIE_INCLUSION_STATUS events for details about each candidate cookie that was considered for sending or setting on a given URL Request. In particular, watch for cookies that were excluded due to SameSite or similar problems.

Scenario: HTTPS Handshaking Issues

Look for SSL_CONNECT_JOB entries. Look at the raw TLS messages on the SOCKET entries. Read the HTTPS Certificate Issues section next.

Scenario: HTTPS Certificate Issues

Note: While NetLogs are great for capturing certs, you can also get the site’s certificate from the browser’s certificate error page.

When a HTTPS connection is established, the server sends one or more certificates representing the “end-entity” (web server) and (optionally) intermediate certificates that chain back to a trusted CA. The browser must use these certificates to try to build a “chain” back to a certificate “root” in the browser/OS trust store. This process of chain building is very complicated.

NetLogs include both the certificates explicitly sent by the server as well as the full chain of certificates it selected when building a chain back to the trusted root (if any).

The NetLog includes the certificates received for each HTTPS connection in the base64-encoded SSL_CERTIFICATES_RECEIVED events, and you can look at CERT_VERIFIER_JOB entries to see Chromium’s analysis of the trust chain. If you see ocsp_response entries, it means that the server stapled OCSP responses on the connection. is_issued_by_known_root indicates whether the chain terminates in a “public” PKI root (e.g. a public CA); if it’s false it means the trust root was a private CA installed on the PC. A verification result of Invalid means that the platform certificate verifier (on Windows, CAPI2) returned an error that didn’t map to one of the Chromium error statuses.

To view the certificates, just copy paste each certificate (including ---BEGIN to END---) out to a text file, name it log.cer, and use the OS certificate viewer to view it. Or you can copy the certificate blocks to your clipboard, and in Fiddler, click Edit > Paste As Sessions, select the Certificates Inspector and press its Content Certificates button.

If you’ve got Chromium’s repo, you can instead use the script at \src\net\tools\print_certificates.py to decode the certificates. There’s also a cert_verify_tool in the Chromium source you might build and try. For Mac, using verify-cert to check the cert and dump-trust-settings to check the state of the Root Trust Store might be useful.

In some cases, running the certificate through an analyzer like https://crt.sh/lintcert can flag relevant problems.

Scenario: Authentication Issues

Look for HTTP_AUTH_CONTROLLER events, and responses with the status codes 401, 403, and 407.

  • If the HTTP_AUTH_CONTROLLER does not show AUTH_HANDLER_INIT, it indicates that the handler did not initialize, for instance because the scheme was unknown or disabled.
  • You might find that the authentication fails with ERR_INVALID_AUTH_CREDENTIALS unless you enable the browser’s DisableAuthNegotiateCnameLookup policy (Kerberos has long been very tricky).
  • You can look at the parameters on the AUTH_LIBRARY_ACQUIRE_CREDS event to understand whether ambient credentials were tried, or whether an explicit username and domain were provided via the credential prompt.
  • If a SSL_CONNECT_JOB or SOCKET events’ Description field starts with pm/, it means that the connection is in ‘privacy mode’ to be used for CORS no-credentials requests, as is used for link fetches that specify the crossorigin attribute.

Scenario: Debugging Proxy Configuration Issues

See Debugging the behavior of Proxy Configuration Scripts.

Got a great NetLog debugging tip I should include here? Please leave a comment and teach me!


Debugging Proxy Configuration Scripts in the new Edge

I’ve written about Browser Proxy Configuration a few times over the years, and I’m delighted that Chromium has accurate & up-to-date documentation for its proxy support.

One thing I’d like to call out is that Microsoft Edge’s new Chromium foundation introduces a convenient new debugging feature for debugging the behavior of Proxy AutoConfiguration (PAC) scripts.

To use it, simply add alert() calls to your PAC script, like so:

alert("!!!!!!!!! PAC script start parse !!!!!!!!");
function FindProxyForURL(url, host) {
alert("Got request for (" + url+ " with host: " + host + ")");
return "PROXY";
alert("!!!!!!!!! PAC script done parse !!!!!!!!");

Then, collect a NetLog trace from the browser:

msedge.exe --log-net-log=C:\temp\logFull.json --net-log-capture-mode=IncludeSocketBytes

…and reproduce the problem.

Save the NetLog JSON file and reload it into the NetLog viewer. Search in the Events tab for PAC_JAVASCRIPT_ALERT events:

Even without adding new alert() calls, you can also look for HTTP_STREAM_JOB_CONTROLLER_PROXY_SERVER_RESOLVED events to see what proxy the proxy resolution process determined should be used.

One current limitation of the current logging is that if the V8 Proxy Resolver process…

… crashes (e.g. because Citrix injected a DLL into it), there’s no mention of that crash in the NetLog; it will just show DIRECT. Until the logging is enhanced, users can hit SHIFT+ESC to launch the browser’s task manager and check to see whether the utility process is alive.

Try using the System Resolver

In some cases (e.g. when using DirectAccess), you might want to try using Windows’ proxy resolution code rather than the code within Chromium.

The --winhttp-proxy-resolver command line argument will direct Chrome/Edge to call out to Windows’ WinHTTP Proxy Service for PAC processing.

Differences in WPAD/PAC Processing

  • The WinHTTP Proxy Service caches proxy authentication credentials (in CredMan) and automatically reuses them across apps and browser launches; Chromium does not.
  • The WinHTTP Proxy Service caches WPAD determination across process launches. Chromium does not and will need to redetect the proxy each time the browser reopens.
  • Internet Explorer/WinINET/Edge Legacy call the PAC script’s FindProxyForURLEx function (introduced to unlock IPv6 support), if present, and FindProxyForURL if not.
  • Chrome/Edge/Firefox only call the FindProxyForURL function and do not call the Ex version.
  • Internet Explorer/WinINET/Edge Legacy expose a getClientVersion API that is not defined in other PAC environments.
  • Chrome/Edge may return different results than IE/WinINET/EdgeLegacy from the myIpAddress function when connected to a VPN.
  • Edge 79+/Chrome do not allow loading a PAC script from a file:// URI. (IE allowed this long ago, but it hasn’t been supported in a long time). If you want to try testing a PAC script locally in Chromium-based browsers, you can encode the whole script into a DATA URL and use that. IE/Edge Legacy do not support this mechanism.

Notes for Other Browsers

  • Prior to Windows 8, IE showed PAC alert() notices in a modal dialog box. It no longer does so and alert() is a no-op.
  • Firefox shows alert() messages in the Browser Console (hit Ctrl+Shift+J); note that Firefox’s Browser Console is not the Web Console where web pages’ console.log statements are shown.


Celebrating Fifteen Years

While lately I’ve been endlessly streaming the latest news with horrified fascination, this morning my calendar unexpectedly popped up a reminder set over a year ago… Today is the fifteenth anniversary of my big-league blogging debut on the Internet Explorer Team’s blog.

My first post there, “A HTTP Detective Story” remains one of my favorites. Sadly, its topic feels all too familiar: a website took a dependency upon a browser quirk, the Referer and User-Agent headers were critical elements of the repro, and I used Fiddler to root-cause the problem. I’ve learned (and shared) so much more about these topics over the last fifteen years, and I appreciate the readers who’ve followed me from the IEBlog to IEInternals (237 posts) to Telerik’s Fiddler blog (41 posts) to my book, to this, my newest blog, TextSlashPlain (189 posts and counting!).

Here’s to learning and sharing for the next 15 years!

With gratitude,