Author Archives: jmacy

API Testing Pillars

As APIs have led the way to IT modernization and are the new fabric of IT infrastructure, active and aggressive API testing has become crucial. Comprehensive Functional, Performance, Interoperability and Vulnerability Testing form the Pillars of API Testing. Only by … Read More

Hidden Costs of Open Source API Testing Tools

In this article, I will discuss some considerations and potential limitations of adopting an open source API testing solution for a long term API strategy . Open Source has become an essential and popular resource for many tools and platforms … Read More

Intro to MTOM

MTOM OVERVIEW With web services-based SOA deployed across Global 2000 enterprises, transmitting attachments such as MRI Scans, X-Rays, Design Documents and Business Contracts using SOAP messages has become a common practice. SOAP Message Transmission Optimization Mechanism (MTOM), is a W3C … Read More


Identity Management plays a pivotal role in securing API-based applications on premise and in the cloud.    Enterprises are under unrelenting pressure to streamline business processes by integrating internal systems and external trading partners. As enterprises integrate applications internally and across corporate boundaries, identity management systems are being tasked to extend identities, roles and access controls seamlessly across internal and external applications.
The need for flexible identity management is primarily driven by rapid API proliferation and security standards adoption. API security  has introduced message-based identity representation and use. Such message-based identities can easily traverse corporate and system boundaries. These Web Services Security standards that promise relief for API applications developers concerned about security, result in grief for IT personnel concerned about testing identities.  Without guidance and tools that test API Identity profiles, this grief in turn results in Identity Fatigue Syndrome.

Symptoms of Identity Fatigue Syndrome

Identity Management solutions already exist in the User-to-Application space and provide robust Single Sign On Solutions (SSO) that enable users to access various business applications with a single identity token. The following are key characteristics of identities in the User-to- Application domain:

  • Transport-level credentials: Typical user credentials in SSO include HTTP Basic Auth, SSL Auth, Two Factor Auth, and Kerberos.
  • Authentication: Prove user identity based on supplied credentials
  • Authorization: Control access to resources, typically URIs, based on user credentials

These characteristics of the identity management create a burden in successfully integrating browser- based clients securely with business applications since there are no standard tools that test all aspect of user-based identities for access control permissions. Instead of providing comprehensive testing tools for deploying Identity management systems, identity product vendors provide rudimentary trouble-shooting guides on how to diagnose access characteristics. For example, if an authorized user is denied access to a URI, detailed access-log and configuration audit-log analysis reveals improper memberships or resource restrictions. At best, identity vendors supply the customer with ad-hoc tools and manual methodologies to test pre-deployment access control functionality. The identity-testing crisis is further exacerbated by testing tools and testing methodologies lagging behind new identity product releases.
Today, the API paradigm based on Web Services has ushered in an era of loosely coupled and chained API-enabled applications [1]. To integrate these applications securely and with minimum friction, new API Security standards have been introduced. These new standards enable existing transport-based identities (HTTP Basic Auth, SSL X509s, Kerberos) to traverse application and corporate boundaries. The API security standards have increased the identity testing challenge because of several factors:

  • API security standards support enveloping of exiting and new identities such as, Username tokens, X.509 certificates, Kerberos tickets and SAML or OAuth assertions using flexible message-based formats.
  • The above identities can appear in XML messages over diverse protocols such as
    • HTTP, HTTPS, JMS, and AMQP.
  • Critical business functions exposed by Web Services require fine-grained access control (authentication, authorization) based on these identities
  • Compliance and regulatory pressures mandate that all proper access controls are in place across the enterprise applications.

The above factors increase the workload on developers and QA personnel, already inundated with developing, testing and deploying API-based applications.

Remedy through Identity Testing

There is no silver bullet to eliminate the challenge of testing the numerous scenarios across protocol- and message-based identities for an API. However, certain steps can be taken to alleviate the fatigue that can result in dealing with various identity mechanisms associated with APIs. A successful identity-testing plan should:

  • Include an identity diagnostics tool that mimics the real-world API clients.  This  tool should be fully capable to generate all kinds of identities for positive and negative testing. The negative testing can be achieved through mutation of identities and conditional tests.
  • Implement Identity  Management tasks, such as identity generation.  Such tasks should be a part of functional regression testing, performance testing and compliance assessment.
  • Deploy strong reporting functionality for rapidly isolating trouble spots during various stages of identity testing.
  • Include a diagnostics tool that mimics every Web Services client and acts as an independent auditor. The richer the reporting, the easier it is for compliance officer in an enterprise to audit the access controls in place for Web Services.

Figure 1 is an example of how a API tester could utilize automated diagnostics software from a desktop for complex identity management testing.      Figure 1 shows a typical system that involves authentication of a user whose identity traverses across the enterprise applications.

  • In step 1, a user authenticates into a browser-based portal with a built-in Web Services client.
  • In step 2, after the user is validated, based on his or her credentials, a subsequent SAML assertion is generated that is then signed by the Web Services SAML (Security Assertion Markup Language) is an XML based security token standard that carries authentication and authorization information [4]. SAML is used to envelope other identities and project them across application and enterprise boundaries.
  • The Web Services client then takes two a It sends the requests to the back end Web Services provider, as shown in step 2, while also logging the SOAP message to a central repository as shown by step 3.
  • In step 4, the software diagnostics running on the desktop pulls all the SOAP messages from the repository through a batch It then applies its two regression policies, Automated Positive Regression and Automated Negative Regression on the batched SOAP requests. Both these policies are invoked on the batched SOAP messages through automated tasks (validate SAML, transform, send, analyze). The positive and negative test suites are launched against the target Web Service as shown in step 5.
  • The resultant requests are collected for analysis as shown in step 6. Such tests ensure the reliability and robustness of a back-end API Provider as to how it handles API access controls based on complex identities.  This is an example of how diagnostic software should handle identity testing and alleviate the tester from manually testing various identity combinations.


Identity management systems remain tightly coupled with Web Services deployment to ensure tight access control based on multiple identities. As Web Services applications move towards external partner integration, testing of identities becomes even a more complex task. Robust and reliable access control is only possible through automated regression of identity testing at every point in the Web Services chain. API managers are responsible for providing comprehensive software diagnostics tools and methodologies early in a API projects development life cycle. By deploying such tools and techniques, API Managers ensure that their Web Services-enabled applications are less prone to information theft through access control breaches.