Identity Management in the Cloud

Security - Other
  • Smaller Small Medium Big Bigger
  • Default Helvetica Segoe Georgia Times

What does a business need to store and use its identity data in the cloud?

Editor’ note: This chapter is excerpted from chapter 5 of Identity Management: A Business Perspective (MC Press, 2016).

When it comes to supporting cloud services with identity data, two main questions must be answered: what data repository should we use, and what interfaces are required?

Data Repository

The choice of a data repository will be guided by the required environment. This includes factors such as the predominant supplier for on-premises systems. If an organization is primarily a Microsoft environment, the cloud service environment will likely be Azure. If it’s predominantly a Linux environment, an Amazon Web Services (AWS) or Google environment might be preferred. Table 1 lists some common cloud environments.

Table 1: Cloud environments

Cloud Environment


Microsoft Azure

If an organization maintains a predominantly Microsoft environment, there would need to be a good reason not to deploy an Azure AD environment. Azure AD integrates well with on-premises AD infrastructure, supports all the required protocols, and will provide a degree of “future proofing.” Azure AD is built on a graph database, so it can well accommodate relationship data, which is becoming increasingly important in the identity management space. Finally, the Azure environment supports all the main interface protocols.

AWS Solutions

If an organization has already adopted a SaaS solution provider such as Okta, OneLogin, or PingOne, the cloud identity provider (IdP) solution is already in place. It is most important that any new SaaS app that is engaged should use the existing IdP service, even if the application is not being enrolled in the SaaS solution. For instance, if an organization is a Salesforce customer, consideration should be given to using the Salesforce IdP service for all cloud apps. Regardless of which option is selected, it is important that organizations retain control of their identity repository in the cloud. Additionally, the organization should consider not engaging any SaaS apps that do not support a central identity provider service model.

Google Solutions

If an organization is a Google customer that uses Gmail, Google Docs, and other Google services, the obvious identity management solution is to use Google’s identity and access management service to the degree possible (note: this is not the Google ID public IdP service). Organizations would still undertake their same registration processes but then store the data in their Google cloud service. The benefit is the well-managed API that Google maintains. Support for Google Groups, wide protocol support, and support for SaaS providers such as Okta, OneLogin, or PingOne, is provided.

Private cloud

There are a number of vendors that provide IDaaS services. For instance, an HP customer might engage the HP Enterprise cloud environment, a Dell customer might choose the Dell Cloud Computing Service, or an IBM customer might migrate to IBM cloud services. But once a cloud-based IdP has been deployed, integration with the organization’s identity service should be a prerequisite for selection of a SaaS app.


Developing a Cloud Strategy

As with most IT deployments, it is important to set the strategy for the company before engaging a CSP. There are several components to this, each of which will affect the company’s approach to identity and access management.


If there is one shortcoming in the deployment of many cloud services, it is in the policy area. In many cases, company boards of management are abrogating their corporate responsibilities by not mandating policy in the deployment of their information technology. Too often important decisions are being put in the “too hard” basket or ignored because the board members don’t understand the ramifications. It is important for the CIO to understand the technology and not rely on IT staff to make decisions that, in effect, set corporate policy. The “don’t tell me so I can’t be blamed” approach is not tenable because of the risk of getting corporations into unnecessary strife. In fact, business schools should teach IT basics to business managers to equip them to ask the right questions and evaluate the risks associated with the main approaches to technology deployments.

It’s highly recommended that an enterprise architecture approach be taken. (For more information, see the Enterprise Architecture section in Identity Management: A Business Perspective, page 102). The cloud environment should be subjected to the same rigor that governs the on-premises environment. At the technical level, the operating systems and currently supported versions should be dictated. At the application level, the cloud-based programs should be included in the organization’s application portfolio, along with the on-premises programs. At the information architecture level, entity relationship diagrams should map the data being passed to and from cloud applications and databases in the same way that on-premises applications are documented. The way in which the organization’s (identity provider) IdP interfaces to applications should be included in the enterprise architecture documentation. At the business system architecture level, the focus should be on how cloud-based apps support business systems within the organization, ensuring that the performance of cloud apps meets business unit expectations.

Organizational Capability

A fundamental requirement in any strategy development is understanding the organization’s capability. If the capability to implement a strategy doesn’t exist, or is insufficient, it must either be developed or the strategy changed. As a simplistic example: if a company does not have the technical skills to manage servers and deploy applications, an IaaS solution is not a good idea.

When it comes to identity management, an organization can outsource the management of the IdM infrastructure but cannot outsource the responsibility for it. Regardless of the deployment model selected, it is the company’s responsibility to ensure that it meets regulatory requirements.

User Requirements

It is equally necessary to understand user requirements, and it is here that many organizations are found wanting. IT departments are notoriously bad at getting close to the business units and working with them to design solutions that suit the business. But doing so is mandatory in order to deploy an efficient and effective cloud-based service. To fail to do this will mean that business units will seek out or develop solutions on their own, which will circumvent corporate governance processes and could result in the organization being exposed to unnecessary risk.

A better solution is for the IT department to facilitate connections to SaaS applications, to ensure that an optimal solution architecture is developed and that business functionality is provided within corporate governance guidelines. While business units tend to balk at the cost of governance processes, adhering to governance policy will normally be the least expensive option in the long term.

Data Storage Considerations

There are multiple policy areas that will impinge on storing data in the cloud. Companies must understand that controls are required for the data they will be collecting or retaining. Table 2 lists commonly used data controls.

Table 2: Controls for storing data in the cloud



Personally Identifiable Information (PII)

Most jurisdictions have some sort of privacy legislation in place. Some countries such as the European Union, Singapore, Australia, and Hong Kong have stringent laws in place that dictate how identity data is to be treated and limit the ways in which it can be stored. Adhering to these requirements will ensure adequate protection of identity data and will avoid wasting money on litigation defense.

Payment Card Industry (PCI)

There are now stringent rules on the storage of credit card and banking information on staff and customers. Storage of this sensitive information requires adherence to these rules, and audits need to be performed. It is generally better to not store such information if not necessary. For instance, if transactions are typically infrequent, capturing credit card details to improve users’ experience entails more risk than is warranted.

Corporate-controlled information

Determining the classification of corporate data that can be stored in the cloud is a prerequisite for migrating to the cloud. Doing so will allow the proper determination of other types of cloud services to be engaged and what service level agreements (SLAs) are required. Corporate-controlled information typically includes documents that must be shared for board-level meetings, upper management meeting minutes, product development data, project team documents, production data, and HR information.