Shibboleth Origin Deployment Guide
Shibboleth Origin Deployment Guide
draft-internet2-mace-shibboleth-shib-origin-deploy-19.html
Nate Klingenstein
26 November, 2002
Comments should be directed to .
Shibboleth v0.7 has some limitations and lacks certain
security provisions which will be present in the final version.
It is strongly advised that this version not be used to protect
any sensitive data. Some sections of the deploy guide have not
yet been populated with text. This document describes
additional functionality which will be present in the final
version, but which is not implemented in the v0.7, including
but not limited to:
-
Some security features, including authentication of
SHARs, and support for non-trivial certificate validation
processes beyond simple CA validation.
-
Bundling of SSO or WebISO systems.
-
Any support for customized target side AAP's, except by
writing an Apache module to implement the desired policy.
As shipped, EPPN and eduPersonAffiliation are only accepted
from the AA if the scope value is equal to the value of
"domain" in the origin site HS's web.xml file, which will
be the same as the name provided in registration with Club
Shib.
-
Club Shib is only a rudimentary concept at this time,
and no support for multiple clubs or club policies is
provided. Registration with the Shibboleth project
management team is required to interoperate with other test
sites, unless peer to peer sharing of certificate
information is undertaken. For more information about
acquiring certificates accepted by Club Shib, please refer
to section 2.d.
Functionality which has been added since the previous
version (alpha-2.5) includes:
- A completely new target side implementation. New features
include separation of the SHAR and SHIRE, elimination of the
need for Tomcat unless running a WAYF as well, and dynamic
generation of error pages. Configuration and functionality
are similar to what was found in alpha-2.5.
- A new C/C++ based implementation of OpenSAML.
- Shibboleth and OpenSAML can now optionally be installed
by building from source; however, this is not yet fully
documented.
- The deployment of ARP's will change dramatically before
the next release, rendering ARP's from v0.7
incompatible.
Before starting, please sign up for all applicable
mailing lists. Announcements pertinent to Shibboleth
deployments and developments and resources for deployment
assistance can be found here.
Please send any questions, concerns, or eventual confusion
to .
This should include, but not be limited to, questions about the
documentation, undocumented problems, installation or
operational issues, and anything else that arises. Please
ensure that you have the appropriate
.tarball for your operating system. Thank you for your help
in testing Shibboleth.
Shibboleth Origin -- Table of
Contents
-
- Origin
- Target
- WAYF
- Clubs
-
- Requirements
- Join a
Club
- Security
Considerations
- Server
Certs
- Attribute Release
Policies
- Designate
Contacts
- Browser
Requirements
- Clocks
- Other
Considerations
-
- Software
Requirements
- Deploy HS and
AA
- Implement a MySQL
directory (optional)
-
- Basic
Configuration
-
Key Generation and
Certificate Installation
- Sharing
certificate/key pairs between Apache and Java
keystores (optional)
- Linking the
Authentication System to the HS
- Deploying AA
plug-ins for attributes(Java API)
- Establishing
default ARP's for the origin community
- MyAA
-
-
ARP
Syntax
- Site
ARP
- User
ARP's
- ArpUtil
- MyAA
-
- Basic
Testing
- Logging
- Common
Problems
License Information
Before proceeding with any installation of,
implementation of, or any other use of Shibboleth or its code,
read and agree to the usage terms put forth in the LICENSE file
included in the tarballs. Note that Shibboleth is based on the
Security Assertion Markup Language (SAML), a proposed standard
in the OASIS
organization. There are intellectual property claims on SAML
technology that are published on the OASIS
site. Shibboleth deployers are encouraged to evaluate these
claims and respond to them as they see fit.
1. Shibboleth Overview
Shibboleth is a system designed to exchange attributes
across realms for the primary purpose of authorization. It
provides a secure framework for one organization to transmit
attributes about a web-browsing individual across security
domains to another institution. In the primary usage case, when
a user attempts to access a resource at a remote domain, the
user's own home security domain can send certain information
about that user to the target site in a trusted exchange. These
attributes can then be used by the resource to help determine
whether to grant the user access to the resource. The user may
have the ability to decide whether to release specific
attributes to certain sites by specifying personal Attribute
Release Policies (ARP's), effectively preserving privacy while
still granting access based on trusted information.
When a user first tries to access a resource protected by
Shibboleth, they are redirected to a service which asks the
user to specify the organization from which they want to
authenticate. If the user has not yet locally authenticated to
a WebISO service, the user will then be redirected to their
home institution's authentication system. After the user
authenticates, the Shibboleth components at the local
institution will generate a temporary reference to the user,
known as a handle, for the individual and send this to the
target site. The target site can then use the handle to ask for
attributes about this individual. Based on these attributes,
the target can decide whether or not to grant access to the
resource. The user may then be allowed to access the requested
materials.
There are several controls on privacy in Shibboleth, and
mechanisms are provided to allow users to determine exactly
which information about them is released. A user's actual
identity isn't necessary for many access control decisions, so
privacy often is needlessly compromised. Instead, the resource
often utilizes other attributes that are after they are
associated with an identity, such as faculty member or member
of a certain class. Shibboleth provides a way to mutually refer
to the same principal without revealing that principal's
identity. Because the user is initially known to the target
site only by a randomly generated temporary handle, if
sufficient, the target site might know no more about the user
than that the user is a member of the origin organization. This
handle should never be used to decide whether or not to grant
access, and is intended only as a temporary reference for
requesting attributes.
1.a. Origin
There are four primary components to the origin side in
Shibboleth: the Attribute Authority (AA), the Handle Service
(HS), the directory service, and the local sign-on system
(SSO). The AA and HS are provided with Shibboleth, and an
open-source WebISO solution produced by the University of
Washington known as Pubcookie is also supplied; the directory
is provided by the origin site. Shibboleth is able to
interface with a directory exporting an LDAP interface or a
SQL database containing user attributes, and is designed such
that programming interfaces to other repositories should be
readily implemented. Shibboleth relies on standard web server
mechanisms to trigger local authentication. A .htaccess file
can be easily used to trigger either the local WebISO system
or the web server's own Basic Auth mechanism, which will
likely utilize an enterprise authentication system, such as
Kerberos.
From the origin site's point of view, the first contact
will be the redirection of a user to the handle service,
which will then consult the SSO system to determine whether
the user has already been authenticated. If not, then the
browser user will be asked to authenticate, and then sent
back to the target URL with a handle bundled in an attribute
assertion. Next, a request from the Shibboleth Attribute
Requester (SHAR) will arrive at the AA which will include the
previously mentioned handle. The AA then consults the ARP's
for the directory entry corresponding to the handle, queries
the directory for these attributes, and releases to the SHAR
all attributes the SHAR is entitled to know about that
user.
1.b. Target
There are three primary components to the target side in
Shibboleth: the Shibboleth Indexical Reference Establisher
(SHIRE), the Shibboleth Attribute Requester (SHAR), and the
resource manager (RM). An implementation of each of these is
included in the standard Shibboleth distribution. These
components are intended to run on the same web server.
From the target's point of view, a browser will hit the RM
with a request for a Shibboleth-protected resource. The RM
then allows the SHIRE to step in, which will use the WAYF to
acquire the name of a handle service to ask about the user.
The handle service (HS) will then reply with a SAML
authentication assertion containing a handle, which the SHIRE
then hands off to the SHAR. The SHAR uses the handle and the
supplied address of the corresponding attribute authority
(AA) to request all attributes it is allowed to know about
the handle. The SHAR performs some basic validation and
analysis based on attribute acceptance policies (AAP's).
These attributes are then handed off to the RM, which is
responsible for using these attributes to decide whether to
grant access.
1.c. Where are you from? (WAYF)
The WAYF service can be either outsourced and operated by
a club or deployed as part of the SHIRE. It is responsible
for allowing a user to associate themself with an institution
of their specification, then redirecting the user to the
known address for the handle service of that institution.
1.d. Clubs
A Shibboleth club provides part of the underlying trust
required for function of the Shibboleth architecture. A club
is a group of organizations(universities, corporations,
content providers, etc.) who agree to exchange attributes
using the SAML/Shibboleth protocols and abide by a common set
of policies and practices. In so doing, they must implicitly
or explicitly agree to a common set of guidelines. Joining a
club is not explicitly necessary for operation of Shibboleth,
but it dramatically expands the number of targets and origins
that can interact without defining bilateral agreements
between all these parties.
A club can be created in a variety of formats and trust
models, but must provide a certain set of services to club
members. It needs to supply a registry to process
applications to the club and distribute membership
information to the origin and target sites. This must include
distribution of the PKI components necessary for trust
between origins and targets. There also needs to be a set of
agreements and best practices defined by the club governing
the exchange, use, and population of attributes before and
after transit, and there should be a way to find information
on local authentication and authorization practices for club
members.
2. Planning
There are several essential elements that must be present in
the environment to ensure Shibboleth functions well, both
political and technical. Shibboleth is primarily written in
Java on the origin side. These are the recommendations and
requirements for a successful Shibboleth implementation.
2.a. Requirements
-
A common institutional directory service should be
operational; Shibboleth comes with LDAP and MySQL abilities
built in, and the Attribute Authority has a Java API which
will allow specification of interfaces with legacy
directories. This is discussed further in section 4.d.
-
A method to authenticate browser users must be in place,
preferably in the form of an enterprise authentication
service. Some form of an SSO or a WebISO service is not
explicitly necessary for Shibboleth; however, without it,
users will have to repeatedly authenticate to the home
organization for each new target application domain they
wish to visit. Implementation details of this are discussed
in section 4.c.
-
Shibboleth currently only supports Linux and
Solaris.
-
A web server must be deployed that can host Java
servlets, Tomcat, and optionally a MySQL database(which may
or may not be on the same server as the Shibboleth
components).
2.b. Join a Club
While it is not necessary for a target or origin to join a
club, doing so greatly facilitates the implementation of
multilateral trust relationships. Each club will have a
different application process.
To join Club Shib for the Alpha 2 test period, please
containing the following information:
- Domain Name of the origin site (e.g., Ohio State's is
"osu.edu")
- Complete URL to access the HS
- The CN (usually the hostname) of the HS's certificate's
subject
- Any shorthand aliases the WAYF should support for the
origin site (e.g., Ohio State, OSU, Buckeyes)
- Club contact names and addresses
To interoperate with other sites in Club Shib, the HS will
need to have a private key and associated certificate
generated. When generating the certificate, the subject field
will contain a CN attribute. Often, this will be the hostname
of your Handle Service, particularly if the same key-pair and
certificate will be used for SSL as well. While any name may
be assigned that is acceptible to the signer of your
certificate, using the hostname is strongly encouraged.
If, for some reason, the HS's URL is not yet known, but
its hostname and CN have been determined, the URL may be
supplied later. In the meantime, the WAYF will be unable to
direct users to that HS, but any assertions from the site
will still be accepted by club SHIRE's. When the site is
accepted into the Club, its information is added to the sites
file used by the WAYF and target sites.
For more information on Clubs, refer to 1.d or the Shibboleth v1.0 architectural
document.
2.c. Security Considerations
Shibboleth's protocols and software have been extensively
engineered to provide protection against many attacks.
However, the most secure protocol can be compromised if it is
placed in an insecure environment. To ensure Shibboleth is as
secure as possible, there are several recommended security
precautions which should be in place at local sites.
-
SSL use is optional for origin sites. Club guidelines
should be considered when determining whether to
implement SSL, and, in general, SSL should be used for
interactions with client machines to provide the
necessary authentication and encryption to ensure
protection from man-in-the-middle attacks. It is strongly
suggested that all password traffic or similarly
sensitive data should be SSL-protected. Assessment of the
risk tradeoff against possible performance degradation
should be performed for all applications.
-
Many other attacks can be made on the several
redirection steps that Shibboleth takes to complete
attribute transfer. The best protection against this is
safeguarding the WAYF service and ensuring that rogue
targets and origins are not used, generally by
development of the trust model underneath Shibboleth.
Shibboleth also leverages DNS for security, which is not
uncommon, but attacks concerning bad domain information
should be considered.
-
Information regarding origin users is generally
provided by the authoritative enterprise directory, and
the acceptance of requests from target applications can
be carefully restricted to ensure that all requests the
SHAR performs are authorized and all information the
origin provides is accurate. Proper security measures
should also be in place on directory access and
population(see
Access Control in the LDAP
recipe for more information). Use of plaintext
passwords is strongly advised against.
-
Server platforms should be properly secured,
commensurate with the level that would be expected for a
campus' other security services, and cookie stores on
client machines should be well protected.
2.d. Server Certs
In the Shibboleth architecture, the SHIRE, SHAR, HS, and
AA must all have various client and/or server certificates
for use in signing assertions and creating SSL channels.
These should be issued by a commonly accepted CA, which may
be stipulated by some Club rules. For the Shibboleth Alpha 2
testing, the following CA's will be recognized by Club
Shib:
* The certificates issued by these CA's will expire
fairly quickly and should only be used for testing.
OSU will also provide a test CA to be used during
Shibboleth development. Thawte presently issues
certificates with extKeyUsage restrictions that make them
incompatible with Shibboleth.
2.e. Attribute Release Policies
The Attribute Authority maintains a set of rules called
Attribute Release Policies (ARP's) that define which
attributes are released to which targets. When a browser user
tries to access a resource, the SHAR asks the origin site AA
to release all the attributes it is allowed to know. The SHAR
provides its own name and an optional URL which can further
refine the information the SHAR is allowed to know. The AA
processes this request using all applicable ARP's, determines
which attributes and values it will release, and then obtains
the values actually associated with the browser user. The AA
sends these attributes and values back to the SHAR.
The set of ARP's that is applicable to a given request is
determined by first matching the SHAR name against the ARP's.
After one or more matching ARP's are found, the AA searches
the URL trees pertaining to those ARP's. The closest matches
are selected, and an appropriate list of attributes is
evaluated combining the applicable ARP's with the attribute
generation methods, usually querying the enterprise directory
and applying the appropriate logic and packaging.
An ARP may be thought of as a sort of filter for outbound
attributes; it cannot create attributes or data that weren't
originally present, but it can limit the attributes released
and the values those attributes may have when released. It
does not change the information in the data sources in any
way.
This information is then sent to the requesting SHAR.
Although an arbitrary number of ARP's may be defined for an
arbitrary number of SHAR's and URL trees within SHAR's, only
the most precisely matching ARP is considered a match, and
then used to determine the release of information to a target
site. If no SHAR matches the query, the default set of
attributes as defined by the AA will be released.
A special type of ARP is the site ARP, which applies to
every user that an AA can vouch for, as opposed to user
ARP's, which apply only to the pertinent user. Site ARP's are
administratively created and maintained, with one set of site
ARP's defined for a given AA. If both a site ARP and a user
ARP are applicable to a particular SHAR/URL combination, then
the attributes released are the union of the two ARP's
necessarily excluding any attributes marked exclude and
necessarily including any attributes marked include.
Site ARP's are formed in much the same way as user ARP's,
with a couple noteworthy differences. Site ARP's may force
the release of attributes or particular values to particular
SHAR/URL combinations, and may also prohibit the release of
attributes or particular values to particular SHAR/URL
combinations. The other primary role of the site ARP is to
define a default release policy that applies to all SHAR's
that do not have an entry in either the applicable site ARP's
or user ARP's. For privacy and security reasons, the default
policy will generally be fairly restrictive. Site ARP's give
administrators very powerful ways to apply trust
relationships with information providers and other targets
across an entire user base.
2.f. Designate Contacts
Since Shibboleth deals both with daily technical and
operational issues and also with contractual issues, a set of
contacts should be set up to support the user base and to
facilitate interactions with other Shibboleth sites and club
members. It is recommended that at least technical and
administrative contacts be designated.
2.g. Browser Requirements
A primary Shibboleth design consideration was to require
very little or no modification to client machines. The only
requirement is that a browser is used which supports cookies,
redirection and SSL. Browser users will have to perform an
additional click to submit the authentication assertion if
_javascript_ is not functional.
2.h. Clocks
NTP should
be run on all web servers. Shibboleth employs a short handle
issuance time to protect against replay attacks. Because of
this, any significant degree of clock skew can hinder the
ability of users to access sites successfully.
2.i. Other Considerations
Especially for higher education, there are a handful of
laws enacted which may have important ramifications on the
disclosure of personal information and attributes. Since
Shibboleth does not necessarily need to transmit identity, it
is an ideal solution for many higher education situations.
Nevertheless, all parties within the United States of America
are strongly advised to consult the Family Educational
Rights and Privacy Act of 1974(FERPA), and all other
relevant state and federal legislation before deploying
Shibboleth.
3. Installation
3.a. Software Requirements
- Apache
1.3.26+
- Apache Ant
1.5+
- Tomcat
3.3.1 Java server
-
Sun JRE 1.3.1 (any
subversion)
Other versions of the JRE are not supported and are
known to cause errors when working with
certificates.
-
mod_jk
You may need to build mod_jk against Apache, which
will generally require GCC or a platform-specific C
compiler.
-
An enterprise authentication mechanism
Ideally, this will be a WebISO or SSO system such as
the University of Washington's Pubcookie
package. The minimal requirement is for the web server
to be able to authenticate browser users and supply
their identity to the Handle Server.
-
An enterprise directory service
Shibboleth currently supports retrieving user
attribute information from either LDAP or MySQL. For testing
purposes, Shibboleth also supports a minimal echo
responder which will always return two pre-defined
attributes.
The origin can be built using a MySQL database to store
handles or an in-memory method to store handles.
Additionally, it's possible to define alternate storage
methods using the API. (optional)
3.b. Deploy HS and AA
-
Ensure you have already obtained the proper .tarball.
-
The archive will expand into a shib/
directory(/usr/local/ recommended).
-
Run the following command to move the Java files into
Tomcat's tree:
cp /usr/local/shib/java/shibboleth.war
/usr/local/tomcat/webapps/
-
Restart Tomcat, which will automatically detect that
there has been a new .war file added. This file will by
default be expanded into
/usr/local/tomcat/webapps/shibboleth.
-
Apache must be told to map the URL's for the
Shibboleth HS and AA to Tomcat. Two popular ways of doing
this are to include the following text directly in
httpd.conf, or to place Include
conf/mod_jk.conf in httpd.conf, and place
the following lines in
/etc/httpd/conf/mod_jk.conf:
--------- begin ---------
<IfModule !mod_jk.c>
LoadModule jk_module libexec/mod_jk.so
</IfModule>
JkWorkersFile
"/usr/local/tomcat/conf/jk/workers.properties"
JkLogFile "/usr/local/apache/logs/mod_jk.log"
JkLogLevel emerg
JkMount /shibboleth/* ajp13
--------- end ---------
-
Modify Tomcat's /conf/server.xml as
follows:
-
Add address="127.0.0.1" inside the
<Ajp12Connector> and
<Ajp13Connector> configuration
elements to prevent off-host access.
-
Add tomcatAuthentication="false" to the
<Ajp13Connector> configuration element
to ensure that the user's identity is passed from
Apache to the servlet environment.
3.c. Implement a MySQL directory (optional)
-
A MySQL database needs to be implemented along with a
JDBC driver MM.MySQL such as
for use by the HS. The driver selected should be unpacked
and the .jar file should be moved to
/usr/local/tomcat/lib/apps/.
-
Using /usr/local/shib/etc/shibdump.sql, a MySQL database must be
created. Copy this file where desired. The default access
for Shibboleth to the database is
shib/shib. This can and should be
changed for security purposes in both the database itself
and web.xml.
-
As admin, run mysql <
shibdump.sql; this may be done locally or on another
machine.
4. Getting Running
4.a. Basic Configuration
The main configuration file for Shibboleth's origin side
is located in
/usr/local/tomcat/webapps/shibboleth/WEB-INF/web.xml.
This file contains configuration information for the origin
side in several sections. The first is a set of options that
must be defined outside the servlet configuration, followed
by configuration information for the HS and the AA. The
configuration must be consistent with values elsewhere in the
deployment, such as the HS' certificate
and with directory access usr/pwd, etc., or access errors
will occur. These are the variables that may be specified for
each component of web.xml:
|
repository = <type>
|
This option must be specified outside the servlet
description, and specifies the method used to store
handles. The two currently valid values are
SQL and MEMORY.
|
MySQL(optional -- These values must be
populated outside the servlet description if repository =
SQL):
|
DBdriver = <driver name>
|
This is the name of the driver that the HS should
use in queries to the MySQL database. For MM.MySQL,
this should be org.gjt.mm.mysql.Driver.
|
|
DBuser = <login>
|
This is the username used to login to the MySQL
database, and must be consistent with one defined in
the database.
|
|
DBpass = <password>
|
This is the password used to login to the MySQL
database, and must be consistent with one defined in
the database.
|
|
DBdomain = <domain>
|
Specifies the location of the MySQL server.
Localhost cannot be used as a value due to
processing by the Tomcat server; even if the database
is hosted locally, the FQDN must be used.
|
HS:
|
domain = <domain name>
|
Specifies the domain in which the HS is located,
e.g. internet2.edu. Used to populate the
Subject NameQualifier in issued attribute
assertions.
|
|
HSname = <domain name>
|
Specifies the machine on which the HS is located,
e.g. shib.internet2.edu
|
|
ticket = <milliseconds>
|
Specifies the duration in milliseconds for which
an issued attribute assertion should be valid;
defaults to 1400000. Refer to club
guidelines for advice in populating this field.
|
|
AAurl = <url>
|
Defines the URL where the AA runs, such as
https://shib.internet2.edu/shibboleth/AA.
|
|
KSpath = <pathname>
|
Defines the pathname to the JKS keystore that is
used by the HS. The effective root of this path is
/usr/local/tomcat/webapps/shibboleth and the
keystore should usually be placed in
WEB-INF/conf(Defaults to
/WEB-INF/conf/keystore.jks)
|
|
KSpass = <password>
|
Specifies the password used to access the JKS
keystore.
|
|
KSkeyalias = <alias>
|
Specifies the alias used to access the HS's
private key entry within the keystore.
|
|
KSkeypass = <password>
|
Specifies the password used to access the HS's
private key entry within the keystore.
|
|
certalias = <alias>
|
This is the alias used to access the certificate
associated with the key used by the HS within the
keystore. Should be identical to
KSkeyalias.
|
|
username = <HTTP Request Header>
|
Specifies the CGI header to pull username from
when assigning handles. If omitted,
REMOTE_USER is assumed.
|
AA:
|
domain = <domain name>
|
Specifies the domain in which the AA is located,
e.g. internet2.edu. This is the default
scope for attributes.
|
|
arpFactoryMethod = <method>
|
This will eventually allow for the selection of
how ARP's are stored, supporting SQL databases and
LDAP repositories. Currently, file is the
only method supported.
|
|
ctxFactoryClass = <parameter>
|
This optional parameter allows sites to override
how AA will populate attribute assertions. This is
only needed if LDAP is not used.
An echo setting is provided for testing by
specifying the parameter
edu.internet2.middleware.shibboleth.aaLocal.EchoCtxFactory
which will automatically return
eduPersonAffiliation=member and
eduPersonPrincipalName populated with the
UID used to login the user in question.
|
LDAP(optional -- These values must be
provided if LDAP is used(e.g. ctxFactoryClass not
specified)):
|
dirUrl = <LDAP URL>
|
This is the URL of the LDAP directory from which
Shibboleth should retrieve user attributes. It should
contain the LDAP hostname and search base. An example
query URL would be
ldap://shib2.internet2.edu/ou=People,dc=internet2,dc=edu
|
|
ctxPrincipal = <DN>
|
This optional parameter allows for specification
of the DN to be used if you want the HS to
BIND to the LDAP server before issuing the query.
|
|
ctxCredentials = <password>
|
This parameter defines the password used
for the LDAP BIND. Must be specified if
ctxPrincipal is populated.
|
|
ldapUserDnPhrase = <DC>
|
This is the prefix for the last part of the DN;
e.g., uid=. The user's ID is dynamically
appended to this value to create the user's complete
DN. Note that if username is not part of the LDAP DN,
a customized ctxFactoryClass must be built
and used for LDAP to work. Additionally, %s
may be specified as part of <DC>; if
%s is specified, it will be replaced with
the user's ID and is used as an LDAP search filter.
The query must return exactly one entry.
|
4.b. Key Generation and Certificate
Installation
The SAML messages generated by the HS must be digitally
signed. Each HS must be issued a private and public keypair,
which is stored in a Java keystore. The current
implementation of Shibboleth requires the use of an ordinary
file-based keystore. The keytool program is included with the
Java development and runtime kits. Access parameters to the
keystore will need to be consistent with those specified in
web.xml.
A sample keystore that can be used is included in the
distribution and can be found in
/usr/local/tomcat/webapps/shibboleth/WEB-INF/conf/keystore.jks
with a password of shibhs.
The following commands will generate a new RSA keypair and
store it in the keystore.jks file, with a keyentry
alias of hs and new passwords of your choosing:
$ cd
/usr/local/tomcat/webapps/shibboleth/WEB-INF/conf
$ keytool -storepasswd -keystore keystore.jks -new
<newpassword>
$ keytool -genkey -keystore keystore.jks -alias hs -keyalg
rsa -keysize 2048
You will be prompted for passwords during key generation
as needed, to access the keystore and assign the key itself
its own password. You will also be prompted for the
distinguished name components to associate with the key. This
DN will be placed in a self-signed certificate and will be
the name that is associated with your HS by Shibboleth. In
particular, the first component you enter for Name will be
the Common Name(when keytool asks for first and last
name, common name is intended), which in most cases should be
the hostname of the HS system. Note that a specific club of
sites may dictate what type of key algorithm, key size, or
validity period is appropriate. For Club Shib, RSA should be
used with a minimum keysize of 2048 bits.
Once you have a keypair generated, the self-signed
certificate must be replaced with a certificate signed by a
CA acceptable to the club you will be joining. If your
certificate is signed by an intermediate CA, such as a campus
CA which has been signed by CREN,
the trust will fail unless the intermediate CA is recognized
as a trusted root by club members as well as the superior CA.
This restriction may change in subsequent releases.
To generate a certificate signing request for a CA, use
the following command:
$ keytool -certreq -keystore keystore.jks -alias hs
-file <csr-file>
The contents of <csr-file> can then be sent
to a CA for signing. You will receive a signed certificate in
return in a file. To install the new certificate into your
keystore, use the following command:
$ keytool -import -keystore keystore.jks -alias hs
-file <cert-file>
Note that if the signing CA's certificate is not already
installed in your keystore as a trusted signer, you may need
to download the CA's root certificate and import it into the
keystore file under a different alias, using a command
similar to the above.
4.b.i. Sharing certificate/key pairs
between Apache and Java keystores (optional)
The JDK includes the command line program
keytool for managing Java keystores. This utility
cannot import or export private key information, making it
difficult to use the same private key and certificate for
Apache and Java-based applications. The Shibboleth
distribution includes extkeytool, a program that
can be used in conjunction with keytool to perform
these tasks. Select the appropriate step-by-step procedure
for your situation from the following guides.
If you have a pre-exiting RSA key/certificate
combination in a keystore and you would like to use it with
Apache:
-
Determine the alias of the keystore keyEntry
containing the key you would like to use in your Apache
setup. Assuming that your keystore is named
yourstore, the following command should
present a list of the entries in the keystore.
$ keytool -list -v -keystore
yourstore
-
Assuming that you identified the appropriate alias
as youralias and the password for the keystore
is yourpass, enter the following command to
export the key in Base64-encoded pkcs8 format.
$ extkeytool -exportkey -keystore yourstore
-alias youralias -storepass yourpass -rfc -file
yourkey.pkcs8
-
In order to use this key with Apache, you must
convert it to PEM-encoded RSA native format. You have
the option of storing the key unencrypted or
encrypted:
-
To use the unencrypted format, enter the
following command for the conversion:
$ openssl pkcs8 -in yourkey.pkcs8
-nocrypt|openssl rsa -out yourkey.key
-
To use the encrypted format, enter the following
command for the conversion:
$ openssl pkcs8 -in yourkey.pkcs8
-nocrypt|openssl rsa -des3 -out
yourkey.enckey
-
The following command will export the corresponding
certificate.
$ keytool -export -keystore yourstore -alias
youralias -rfc -file yourcert
-
Set the mod_ssl
SSLCertificateKeyFile and
SSLCertificateFile directives to point to the
two files you have just created. Take care to remove
any temporary files you created (i.e.
yourkey.pkcs8) and set appropriate file
permissions, especially if you chose to store the key
in an unencrypted format.
If you have a pre-existing RSA key/certificate
combination that you use with Apache and would like to
import it into a java keystore:
-
Convert the private key to unencrypted DER-encoded
pkcs8 format. Assuming your PEM-encoded key is stored
in a file named yourkey.enckey, enter the
following command.
$ openssl pkcs8 -in yourkey.enckey -topk8
-nocrypt -outform DER -out yourkey.der.pkcs8
-
Create a certificate bundle file. This file should
include a series of PEM-encoded X509 certificates
representing a complete trust chain, from the root CA
certificate to the certificate that matches your
private key. If your certificate is stored in a file
named mycert and the CA signer certificate is
stored in a file named ca.cert, you might
enter the following command to create the bundle.
$ cat mycert ca.cert > cert.bundle
Note: mod_ssl-enabled Apache
installations include a number of commonly recognized
CA certificates in the ca-bundle.crt file
under the $ServerRoot/conf/ssl.crt/
directory.
-
Import the key and certificate into the keystore.
Assuming you have already created a keystore named
yourstore with a password of of
yourpass, enter the following command to store
the data under the alias youralias.
$ ./extkeytool -importkey -keystore yourstore
-alias youralias -storepass yourpass -keyfile
yourkey.der.pkcs8 -certfile cert.bundle -provider
org.bouncycastle.jce.provider.BouncyCastleProvider
-
You can verify that the import was successful by
listing entry. Use the command below.
$ keytool -list -v -keystore yourstore -alias
youralias
-
Remember to delete yourkey.der.pkcs8, as it
contains your unencrypted private key.
If you are starting from scratch and do not yet have
a certificate/key pair:
-
Generate an RSA private key. Use the command below,
substituting yourkey with an appropriate name
to use to refer to the key.
$ openssl genrsa -des3 -out yourkey.enckey
1024
-
The following command generates a Certificate
Signing Request, which should be communicated to a
Certificate Authority.
$ openssl req -new -key
yourkey.enckey
-
The Certificate Authority should respond with a
PEM-encoded X509 certificate. Set the mod_ssl
SSLCertificateKeyFile directive to point to
the key file you just created and the
SSLCertificateFile directive to point to file
containing the certificate issued by the Certificate
Authority. Previous sections explaion how to share the
key/certificate pair with a Java keystore.
4.c. Linking the Authentication System
to the HS
The interaction between the HS and the local
authentication system is implemented by supplying the HS with
the identity of the browser user. Most often, this will mean
protecting the HS servlet with some form of local
authentication that populates REMOTE_USER, or
another header specified to the HS via web.xml.
Location blocks can be added to httpd.conf,
associating the appropriate authentication mechanism with the
URL of the HS servlet. The following example demonstrates
association of a very basic authentication method with the
HS:
<Location /shibboleth/HS>
AuthType Basic
AuthName "Internet2 Handle Service"
AuthUserFile /usr/local/apache/conf/user.db
require valid-user
</Location>
Note that .htaccess files cannot be used for this purpose
because URL's are "virtualized" by Tomcat.
It is recommended that the origin be tested at the end of
this process using the process described in section 6.a.
4.d. Deploying AA plug-ins for
attributes(Java API)
Under development and extremely likely to fluctuate with
future implementation.
4.e. Establishing default ARP's for the
origin community
An ARP determines which attributes are released to a SHAR
when a user tries to access a resource. Because of the
architecture of the Shibboleth protocols, the requesting SHAR
is the component that must be considered first when
evaluating which ARP to use. After the requesting SHAR has
been identified, the URL trees defined for that SHAR are
considered, from which the best match is selected and the
specified attributes are released. The data sources for the
AA are authoritative and information they do not contain
cannot be created by use of ARP's in any way.
ARP's can be defined either on a site-wide basis or for
individual users. Although an arbitrary number of ARP's may
be defined for an arbitrary number of SHAR's and URL trees
within SHAR's, only the most precisely matching ARP is used
determine the release of information to a target site. Domain
names are processed first in reverse component order,
followed by the processing of the rest of the URL in normal
order. * is a valid wildcard and all ARP's are
assumed to terminate in a *. Filters can also be
constructed to screen attributes before release to pull out
site-specific information, or release only specific parts of
given attributes.
Site ARP's are defined differently, and only one set of
site ARP's can be defined for a given AA. Several special
types of ARP may be defined in the site policy. One default
ARP for any SHAR or URL that is not matched may be formed.
The release of certain attributes to certain targets may be
marked exclude, which disallows the release of that
attribute, even if the user tries to specify it for themself.
Similarly, certain attributes to certain targets may be
marked include, which disallows the user from
denying the release of that attribute. Filters are evaluated
for both includes and excludes, which screen the released
attributes. A filtered include for a value that a
user does not have will not force the release of that value.
If both a site ARP and a user ARP are applicable to a
particular SHAR/URL combination, then the attributes released
are the union of the two ARP's necessarily excluding any
attributes marked exclude and necessarily including
any attributes marked include.
A set of default ARP's should be established for the
community. When Shibboleth is installed, there are no defined
ARP's, and therefore, nothing will ever be released. This
must be modified using either the text-based editor or the
MyAA webapp. Since these settings will govern the privacy of
your users, site ARP's should be defined carefully before any
sensitive information is utilized, and users should be
informed of their reponsibility to regulate the release of
sensitive information.
Management of ARP's may be delegated to other users to
allow for distributed editing. Each component of each ARP --
SHAR, URL, and attribute -- may be given an ACL. Users listed
in the ACL will be granted the ability to modify that
particular ARP object and create additional ARP objects
further down the ARP tree: the manager of a SHAR ARP could
add new URL trees, but the manager of a URL tree could not
add new URL trees within the SHAR. Control of an ARP object
grants the ability to modify any component on the ARP tree
beneath that object.
For more precise information regarding how ARP's are
processed or syntactically formed, please refer to section 5.a.i.
4.f. MyAA
MyAA is a web-based tool for the management of ARP's. It
is intended primarily for users who want to expand the
default ARP's to handle special needs and applications, or
impose stricter privacy parameters. MyAA's web interface
should be configured and placed as appropriate for your
institution. Users should be made aware of how to best use
the tool for their requirements. Please refer to section 5.a.iv for technical information.
5. Advanced Configuration
5.a. ARP Syntax
This section applies primarily to the syntactic and
technical details of ARP's. For basic information on and
explanation of the construction of ARP's, processing logic,
and ARP management, please refer to section 4.e.
ARP's are stored in the form of several objects: ARP,
ARP:SHAR, ARP:RESOURCE, and ARP:ATTR. Each of these objects
has an associated ACL which defines the set of users who can
define and modify that object and its subordinate objects;
for example, if user parviz is in the ACL of an
ARP:SHAR of shar.osu.edu, then parviz has the
permission to create a new ARP:RESOURCE object hanging from
the shar.osu.edu ARP:SHAR object. This allows for delegated
management of ARP's.
5.a.i. Site ARP
The site ARP must be formed for the ARP user
admin; ARP's for other users are applicable only
to attribute release requests based on that user. If both a
site ARP and a user ARP are applicable to a particular
SHAR/URL combination, then the attributes released are the
union of the two ARP's necessarily excluding any attributes
marked exclude and necessarily including any attributes
marked include. ARP's may be managed both using the ArpUtil
tool, and the MyAA web interface.
Following is an example of a site ARP. The SHAR
no.other.match is defined as the default SHAR,
must have a URL tree of *, and there must be only
one present in any valid site ARP; defaults must not be
used in user ARP's. This SHAR is used if no match is found
for a requesting SHAR in the union of the site ARP and the
applicable user ARP. It is recommended that an extremely
minimal set of information be released to the
no.other.match SHAR. Note that the site ARP is
never evaluated alone, and must always be interpreted in
the context of a particular user. If there is no SHAR/URL
combination specified by the user or site ARP that matches
those in the request, the default is always used.
ARP: admin
SHAR:
no.other.match (default)
URL:
*
ATTR:
eduPersonAffiliation
SHAR:
shar.osu.edu
URL:
http://www.osu.edu/research
ATTR:
eduPersonPrimaryName (include)
ATTR:
eduPersonAffiliation filter: employee
SHAR:
shar.mit.edu
URL:
http://*.mit.edu
ATTR:
eduPersonPrimaryName (exclude)
eduPersonAffiliation will be released to any SHAR that
is not defined in either the admin or applicable user
ARP.
Requests from shar.osu.edu for the
http://www.osu.edu/research tree will receive the
valid eduPersonPrimaryName for any user; users
cannot deny the release of this attribute or any values of
this attribute in their own ARP's.
Employee will always be filtered from release for all
users to shar.osu.edu at
http://www.osu.edu/research, regardless of the
user ARP; however, if and only if that user has some other
value of eduPersonAffiliation defined in the
directory, then that other value will still be
released.
Requests from shar.mit.edu for any URL of the
form http://*.mit.edu/* will never be released
eduPersonPrimaryName. Users cannot release this attribute
to this SHAR/URL combination.
5.a.ii. User ARP's
User ARP's can only block or permit the release of
specific attributes. For example, if a site maintains
attributes A, B, and C, then the user must first define an
ARP for attribute A. Then, the user proceeds to create a
filter, to block the release of certain values of that
attribute. However, a user cannot define an ARP such that
all attributes but A should be released. Additionally,
ARP's that are set as exclude by a site ARP cannot be
modified by the user. The full details of the ARP's that
are applicable to a principle can be discovered using the
ArpUtil tool, although MyAA is more appropriate for
individual users and should provide adequate functionality.
Users may also see notes associated with ARP's, which may
be used for functions such as "We have forced this release
to allow your collaboration in Physics 202 with Brown
University." Additionally, users may create notes to
themselves here when specifying user ARP's.
The following example of the ARP's that would apply to
the user parviz, as obtained using ./ArpUtil
list parviz -acls. Only the user ARP is visible,
although all applicable site ARP's will be evaluated for a
request as well.
ARP: parviz
ACL: arpAcl{[+blk([INSERT]), +dousti([ALL])]}{[]}
SHAR:
shar.cmu.edu
ACL:
sharAcl{[+dousti([ALL])]}{[]}
URL:
http://www.cmu.edu [edu, cmu, www]
ACL:
resourceAcl{[+dousti([ALL])]}{[]}
EPPN
AFFILS
filter: staff, faculty,
URL:
http://www.cs.cmu.edu [edu, cmu, cs, www]
ACL:
resourceAcl{[+dousti([ALL])]}{[]}
EPPN
SHAR:
shar.mit.edu
ACL:
sharAcl{[+dousti([ALL])]}{[]}
URL:
http://*.mit.edu [edu, mit, *]
ACL:
resourceAcl{[+dousti([ALL])]}{[]}
AFFILS
SHAR:
shar.osu.edu
ACL:sharAcl{[+blk([ALL])]}{[]}
URL:
*.edu [edu, *]
ACL:
resourceAcl{[+blk([ALL])]}{[]}
AFFILS
filter: staff, faculty, employee,
dousti has the permission to modify any entry
except for the last SHAR and all information beneath it,
which may be maintained only by blk.
The release of EPPN and
eduPersonAffiliation is allowed to anything within
the SHAR shar.cmu.edu and the URL tree
http://www.cmu.edu/*, and
eduPersonAffiliation values of only staff
or faculty such that if
eduPersonAffiliation contains student,
faculty, employee, then only faculty will be
released; if it is staff, faculty, then the value
released will be staff, faculty; if it is
employee, then the value will be null.
EPPN alone will be released to the URL tree
http://www.cs.cmu.edu if and only if the querying
SHAR is shar.cmu.edu.
Any URL of the form http://*.mit.edu behind the
SHAR shar.mit.edu will be released the full
eduPersonAffiliation of the principle and nothing
more.
Any URL of the form *.edu behind the SHAR
shar.osu.edu will only be released values of
eduPersonAffiliation that are staff,
faculty, or employee.
5.a.iii. ArpUtil
The ArpUtil tool is more powerful than the web interface
and provides functionality the web interface does not.
ArpUtil is part of the distribution and will by default be
in /usr/local/shib/beta/.
These are the commands ArpUtil supports.
|
ArpUtil list <arp name> [-acls]
|
This command will list all the ARP logic
pertaining to the arp name principle.
|
|
ArpUtil add <arp name> [-admin] <shar
name> [-default] <url> <attribute
name> [-exclude] [-filter [!]<val1>
[!]<val2> ...]
|
This command will create a new ARP for a given
SHAR, URL tree, and attribute name, or add to an
existing ARP if specified.
|
|
ArpUtil remove <arp name> [<shar
name> [<url> [<attribute
name>]]]
|
This command will remove a given ARP, or parts
of an ARP instead if specified.
|
|
ArpUtil setAcl <user> <acl> <arp
name> [<shar name> [<url>]]
|
This command will alter the ACL for a given ARP
object.
|
These are the options associated with the various
ArpUtil commands in alphabetical order.
|
acl
|
This specifies the ACL to be applied. ACL's
currently take the form of username.
|
|
-acls
|
Provides a more detailed ARP list, including ACL
information for each object.
|
|
-admin
|
Used to modify the site ARP. Site ARP's are
applicable to all principles referenced by an
AA.
|
|
arp name
|
Provides the a name for the ARP. This can be
arbitrarily specified, but will be used in future
modifications to the ARP.
|
|
attribute name
|
The attribute name for which the ARP is valid.
This must match the representation of the attribute
within the AA logic.
|
|
-default
|
This option negates the SHAR name and makes the
ARP valid for all SHAR's. Valid only for
-admin.
|
|
-exclude
|
This option specifies that this ARP must be
followed and cannot be superceded. Valid only for
-admin.
|
|
-filter
|
This defines the filter that will be applied to
the attribute value when evaluating an ARP. An
attribute value will be permitted for release if it
is val1, etc., but if the ! modifier is specified,
then the filtered value will not be released even
if it is part of the attribute value. See the above examples for use
cases.
|
|
shar name
|
Provides the SHAR with which the ARP should be
associated, e.g. shar.mit.edu.
|
|
url
|
The URL tree for which the ARP is valid. All
*'s are treated as valid wildcards, and
entered URL's are assumed to terminate with
*. Otherwise, the URL must be properly
formed.
|
|
user
|
This specifies which user's ARP objects will be
modified.
|
5.a.iv. MyAA
This section applies primarily to the technical details
of MyAA. For basic information on MyAA, please refer to
section 4.f.
MyAA is intended to allow for a simple, web-based
interface for users to modify their own ARP's. There is
functionality ArpUtil supports that MyAA does not; for
example, SHAR's are masked. Users currently login simply by
entering their name, and the following page will allow
modification of ARP's for the logged-in user. The following
options are available from the webpage once the user is
logged in:
|
Add New Resource
|
This will pull up a new page which will allow a
user to specify a new ARP:RESOURCE within an
existing ARP:SHAR. Resource URL is the URL tree for
which the ARP is valid. All *'s are
treated as valid wildcards, and entered URL's are
assumed to terminate with *. The
attributes that may be released and the values that
this attribute may contain are displayed, and a
checkbox is provided. If the box is checked, the
attribute will be released; if not, it will be
explicitly prohibited from release.
This cannot override existing ARP's that are set
exclude or include.
|
|
Delete Entire ARP
|
This will delete all ARP's associated with a
user that the user is permitted to delete based on
ARP ACL's.
|
|
Delete Resource
|
This will delete an ARP that applies to the user
if the user is permitted to delete this ARP based
on the ARP's ACL.
|
|
Edit Resource
|
This will allow the resource to be edited using
the same interface as is used to create new
resources.
|
6. Troubleshooting
This section provides basic information about testing,
logging, and error handling for Shibboleth origins. This
information is not intended to be comprehensive, but instead
rudimentary guidelines for basic configuration tests and
problems. For more detailed information or answers to specific
problems not addressed in this section, please mail
with a thorough description of errors and configurations
used.
6.a. Basic Testing
Internet2 provides a basic target that can be used to test
origin setup functionality. After your origin is recognized
by Club Shib, simply use any browser to access https://wayf.internet2.edu/shibboleth/sample.jsp.
Select your origin's name and follow the login process as a
user would. Note that SSL must be used, and both the HS and
AA must be fully configured.
The test target will then display a simple page which
includes the basic information sent to it by your origin and
the authentication rules it is using.
For information regarding specific error messages that
may be generated if the origin does not work successfully,
please refer to section 6.c.
6.b. Logging
Shibboleth's origin components log various operations
which may prove useful for auditing, testing, and security
purposes. This data is sent through log4j's standard
mechanism with a log level of INFO. The location of
the log file, the level at which the log is output, the
formatting of the logs, and many more options may be
configured by editing
/WEB-INF/conf/log4j.properties. By default, it is
setup to log to the console of the servlet container, with a
level of WARN, but there is also a commented out
example in the file to give a possible alternate
configuration.
6.c. Common Problems
Shibboleth origins have not yet been sufficiently widely
deployed in diverse environments to identify problems
commonly encountered. At this point, please mail
with any questions or problems encountered for answers to
specific issues. As a knowledge base grows, this section will
be developed.
|
