Protocols

The following network protocols are implemented in StormCrawler:

File

• FileProtocol

HTTP/S

See HTTPProtocol for the effect of metadata content on protocol behaviour.

To change the implementation, add the following lines to your crawler-conf.yaml

  http.protocol.implementation: "org.apache.stormcrawler.protocol.okhttp.HttpProtocol"
  https.protocol.implementation: "org.apache.stormcrawler.protocol.okhttp.HttpProtocol"

• HttpClient
• Selenium
• OKHttp

Feature grid

Features	HTTPClient	OKhttp	Selenium
Basic authentication	Y	Y	N
proxy (w. credentials?)	Y / Y	Y / Y	?
interruptible / trimmable #463	N / Y	Y / Y	Y / N
cookies	Y	Y	N
response headers	Y	Y	N
trust all certificates	N	Y	N
HEAD method	Y	Y	N
POST method	N	Y	N
verbatim response header	Y	Y	N
verbatim request header	N	Y	N
IP address capture	N	Y	N
navigation and javascript	N	N	Y
HTTP/2	N	Y	(Y)
configurable connection pool	N	Y	N

HTTP/2

• the OKHttp protocol supports HTTP/2 if the JDK includes ALPN (Java 9 and upwards or Java 8 builds starting early/mid 2020).
• HttpClient does not yet support HTTP/2
• Selenium: whether HTTP/2 is used or not depends on the used driver

Since #829 the HTTP protocol version used is configurable via http.protocol.versions (see also comments in crawler-default.yaml. Eg., to force that only HTTP/1.1 is used:

http.protocol.versions:
- "http/1.1"