By Ariel Peled and Dr. Troyansky s companies face increasing regulatory requirements such as the Gramm-LeachBliley Act (GLBA) and Sarbanes Oxley Act (SOXA), they must come up with policies and procedures that help them mitigate their liabilities under these regulations. Financial institutions and publicly traded organizations need to prevent the tampering, inappropriate use and leakage of financial, proprietary and non-public information in order to meet these legislative mandates. In this article, we will suggest the top twelve steps that organizations should take to keep non-public information private. We will outline how organizations can establish and enforce organizational policies that will help them comply with recent GLBA and SOXA regulations and reduce their potential legal liabilities. A Introduction Information assets within organizations (for example, financial information used for balance sheets) undergo various stages of processing, validation, reviewing, authorization and assurance until the final version is formed and distributed. Maintaining the confidentiality, integrity and availability of this information along the process poses a non-trivial problem for most organizations, which is nevertheless critical to address not only for regulatory compliance and litigation mitigation, but also for the control and audit over the organization’s most valuable assets. The legal requirements imposed by the GLBA and the SOXA add a sense of urgency. According to the latter act, which was legislated following the “Enron” crisis, officers, top-tier managers and directors may be held personally liable for misdeeds and failures of their subordinates. This legislation narrows the distinction between organizational accountability and personal accountability. Therefore, one must THE ISSA JOURNAL ◆ March 2004 employ the best security practices to mitigate both organizational and personal risks. An information flow process that maintains confidentiality, integrity and availability can be described as a “secure information path.” The following framework outlines a 12-step program that comprehensively addresses the issues of confidentiality, integrity and availability. Step 1: Identify and Classify Information An organization’s first step toward the secure information path is a comprehensive and structured information classification process. The basic methodology for identification and classification is based on defining and treating the information as an asset of the organization— much like a physical asset. Each information asset should be assigned 1) a value, and 2) an information management policy. ▲ Value: A value determines the amount of resources that should be allocated to assure an information asset’s confidentiality, integrity and availability. This value should put a monetary value on the damages, immediate and consequential, that may be incurred by the organization if the asset does not remain confidential, its integrity is breached or it is unavailable. ▲ Information Management Policy: An information management policy determines the asset’s usage and distribution along the information’s lifecycle. Each information asset should have (at least) a confidentiality level (for example, public, confidential, company only, etc.), an integrity level (who is signed on it and when, who changed it since the last signature, etc.) and availability level. By defining these elements in a separate Information Management Policy Database (IMP), the foundation for a secure corporate information infrastructure is established. Step 2: Study Current Processes It is essential to understand current workflows and business processes, both procedurally and in practice, to see how confidential information flows around an organization. The methods for business processes and information flow analysis are known, but they can be quite involved. However, in most cases a short questionnaire that addresses several basic workflow aspects can yield very helpful (and very often surprising) results. Such a questionnaire should identify for each workflow scenario: ▲ The actual participants (agents) ▲ The information assets that are received, created, processed and distributed by each agent ▲ The information flow: the chain of events that consists of the originator, the action, the subject of the action, and the destination entity(s). Step 3: Perform Gap Analysis and Risk Assessment The next step requires a careful analysis of the difference between the defined policies and actual behavior. Disparities by the participants, the information assets in use and the workflows all need to be identified. After the gap analysis is completed, an educated risk assessment process should take place. This assessment includes determining potential breach scenarios, their likelihood, and their potential impact on the organization. ©2004 Technical Enterprises, Inc. Reproduction of this document without permission is prohibited. Typical risk scenarios that should be considered in light of GLBA and SOXA are described below: GLBA: ▲ An employee discloses Non-Public Information (NPI) to an unauthorized recipient by mistake (the “oops…” syndrome). This is perhaps the most common risk, and yet the most dangerous one. ▲ An employee maliciously discloses NPI to an unauthorized recipient. ▲ An employee discloses NPI to an authorized recipient over an unencrypted or otherwise insecure medium, possibly resulting in interception of the NPI by an unauthorized third party. ▲ An employee discloses NPI to a supposedly authorized recipient without proper authorization or authentication. More specifically, the recipient may be an authorized entity but was not properly authenticated as such. ▲ Disclosure or mishandling of account access information or of other NPI results in unauthorized access, change, corruption or destruction of other NPI or to unauthorized transactions and financial damage. SOXA: ▲ An employee changes financial information without proper authorization because the business process in effect differs from the declared procedure. ▲ An employee discloses information regulated by the SEC (such as filings) prior to the release date. ▲ An employee who is required for a financial report production and authorization is left out of the authorization process because of corporate politics. ▲ An employee who is not part of the formal business process participates in financial information gathering for a report because of the employee’s social relations with an authorized employee. ▲ A company regional branch’s actual business process does not execute as defined because a small group of authorized personnel (including the manager) are cooking the books as a result of poor performance. All of the scenarios above should be detected, audited and reported, and eventually mitigated by closing the gap via appropriate policies, procedures and technological means. Because security policy design and implementation may be time consuming, policies need to provide sufficient flexibility for changes in your organization’s business processes. Step 4: Appoint Information Owners and Define Ownership Hierarchy Appointing information owners and defining the ownership hierarchy is absolutely critical for the success of any secure information path project. An information asset to which no clear owner was assigned cannot be properly handled and secured. The owners should have complete responsibility for determining and enforcing information distribution and usage policy on their information assets. Natural candidates for ownership of a certain information asset are either the creator of this information asset or the person who is responsible for it. Examples of these confidential information owners include: ▲ HR director for employee data ▲ CTO for research and patents ▲ CFO for financial reports Nevertheless, various organizations may define different criteria for information ownership. Although in principle the owners have the right and the duty to define and enforce a specific usage and distribution policy (within the limitation of the organizational policies and procedures), the owners are, in many cases, not the information technology (IT) personnel, and they may not be qualified to properly handle confidential information. Therefore, in such cases, a role of custodian of the information is defined. This role is preferably taken by a technical person, for example, a system administrator, IT manager or an information security officer. Just like working a budget, information ownership hierarchy is built from the top down (with bottom-up feedback). First the general hierarchy and framework policies are defined to make sure everyone in the organization shares the same vision and goals. Departmental and regional hierarchies and policies follow, and are derived from the general definition and are confined by it. While different groups in the organization may have different security requirements and needs, a general outline must be followed. Failing to do so may cause, for instance, the marketing department to share customer information with a business partner for the department’s own gain, causing more damage than benefit to the organization as a whole. Step 5: Authorize Content Usage and Processing Policies Users working with confidential or private information need to be authorized, required and limited to perform certain tasks. These tasks can include reading, editing, forwarding, filtering, etc. For instance, some users can read certain documents but do not have the ability to modify them. These user privileges need to be explicitly defined within the content usage policies. In addition, a clear owner should be assigned to each task with reporting procedures and timelines. Ultimately, the privilege to perform these tasks should be monitored using a technology solution that can instantly and clearly report how users are working with sensitive data. This visibility ensures that the organization is in compliance with both internal and regulatory policies. Step 6: Design Access Monitoring and Control Paradigm Constructing a secure information path requires educated and methodological access control. Access control solutions manage the flow of information between entities that request access and entities that contain desired information. Access control serves two main purposes: ▲ Protecting against unauthorized access to organizational resources. ▲ Providing security features that control the manner in which users and systems communicate and interact with other systems and resources. Access control is a well-known practice and is one of the first computer security measures ever employed. Access privileges are typically granted on a file-by-file basis. Yet one often finds that the actual content of a file differs from what it supposed to be. Confidential information can often reside within files that should be restricted. It is therefore imperative to augment file-based access control systems with content-aware technology. ©2004 Technical Enterprises, Inc. Reproduction of this document without permission is prohibited. THE ISSA JOURNAL ◆ March 2004 Step 7: Set Information Distribution Policies should be determined based on the risk analysis performed and the risk level of each node. Setting a solid and manageable information distribution policy is paramount for assuring the required privacy, confidentiality, availability and integrity of your confidential information. The distribution policy imposes a set of limitations on the distribution of an information object, such as the authorized communication channels, authorized senders and recipients, and restrictions and constraints in addition to those with respect to the access and usage of any information asset. In addition, the distribution policy should also include the corresponding set of requirements for action whenever a limitation of restriction is met, such as blocking, alerting, reporting, quarantining or logging. A sound distribution policy often has a high correlation with the file system (or document management app., application services) access policy, and typically may inherit many of its attributes. In general, the policy needs to be defined exclusively, rather than inclusively. Exclusive polices outline which users (or entities) are allowed to access or use which content (an allow list, or “whitelist”). On the other hand, inclusive policies create a restricted list or a “blacklist.” “Blacklisting” may give rise to default grant of access and security risks stemming from classification failure, because each restricted entity needs to be explicitly listed. Step 9: Construct Audit and Detection Program Step 8: Create a Monitoring and Enforcement Plan The ability to monitor and enforce policy adherence will be crucial to the success of the program. Once the policies are defined, it is essential to define and implement monitoring and enforcement points along the path. These control points monitor information access, usage and traffic to verify adherence with the security policy, and perform the actions that the distribution policy defines for compliance and non-compliance. Due to the immense amount of digital information in modern organization workflows, those control points must be both contentaware and have the ability to stop unauthorized operations. Selection and placement of the control points should be placed after each node in the Secure Information Path where major actions are taken (for example, storage devices, applications) or information is transferred in high quantities (for example, e-mail, Web). The number and location of these control points THE ISSA JOURNAL ◆ March 2004 and the envelope. It is therefore preferable to employ a technological means that effectively enforces the retention policy utilizing content aware methods. Step 10: Retention Policy The audit program should cover all aspects of information access, usage and distribution. Detection rules for irregularities should be designed together with actions to take in an event based on its severity. The resulting analysis should be displayed in a clear and concise manner that will highlight any unauthorized or suspicious activity and log authorized actions. Reporting needs to visualize both discrete events and long-term patterns or trends. Three levels of detection are recommended: ▲ Audit reports for authorized actions ▲ Alerts on suspicious activities ▲ Reports regarding unauthorized activity that was blocked New regulations (such as SOXA) place requirements for the duration and manner of confidential information retention. These legal requirements often vary by sector, type of information, type of document and other factors. In addition, it is generally required that the organization formulate and meticulously enforce a policy for the document retention and retrieval. When presented with a subpoena or a request for documents from a regulatory agency or an auditor, it is generally necessary to locate and retrieve all related documentation in a timely fashion. These requests can become formidable tasks if a firm’s retention method relies on tape backup or does not utilize a comprehensive indexing and classification system. In many cases, the content of the document exists in many forms and formats, under various file-names and in various locations (such as the user’s hard-drives). An effective retention policy should thus be defined in terms of the actual content, regardless of the file-name, the format Several laws (such as SOXA) place requirements for the length and manner of retention of information. These requirements are often dependent upon the sector, type of information, type of document, and other factors. In addition, it is generally required that the organization will formulate and meticulously enforce a policy for the retention of documents and document retrieval. When presented with a subpoena or a request for documents from a regulatory agency or an auditor, it is generally necessary to locate and retrieve all related documentation in a timely fashion, which can become an insurmountable task if, for example, the organization’s retention policy consists of a tape backup, or there is no comprehensive indexing and classification system. In addition, it is necessary for the organization to consider the costs and liability that stem from storage of un-required information. It is well known that in the case of litigation or an enquiry (and indeed even before, when the organization has a reasonable expectation to be engaged in litigation or an enquiry), it is necessary to cease the regular cycle of ridding of obsolete documents that may have pertinence to the case in question. A comprehensive retention policy must include policy for handling an upcoming investigation or litigation. In many cases, the content of the document exists in many forms and formats, under various file-names and in various locations (such as the user’s hard-drives) An effective retention policy should thus be defined in terms of the actual content, regardless of the file-name, the format and the envelope. It is therefore preferable to employ a technological means that effectively enforces the retention policy utilizing content aware methods. References ▲ “Harnessing New Regulations for Business Benefit,” Ariel Peled and Lidror Troyansky, www.vidius.com\publications\harnessingregulations.htm. ▲ Handbook of Information Security Management, Micki Krause and Harold F. Tipton, CRC Press LLC, 1998, ISBN: 0849399475. ▲ CISSP All-in-One Exam Guide, Shon Harris and Gareth Hancock, McGraw-Hill Osborne Media; 2nd edition 2003, ISBN: 0072229667. ©2004 Technical Enterprises, Inc. Reproduction of this document without permission is prohibited. Step 11: Plan Data Storage and Availability Knowing where, when and how sensitive information is stored is key to ensuring a secure information path. Storage requirements may include access privileges to storage devices, restrictions regarding portable storage devices (for example, laptop computers), geographic and physical locations, storage space and compression, storage format and encoding, and storage security requirements (for example, encryption, key management, backup and physical security). Step 12: Implement and Deploy New Information Security Policies Management commitment and resource allocation are crucial for the project to succeed. Experience has shown that the risk of project failure and/or opposition from groups within the organization is considerably reduced if you carry out the implementation in the three ‘P’ stages: ▲ ‘Principal’—quick and cheap. Run the process only over information that is critical to the organization or whose security is mandatory by law. For example, financial institutions should start by limiting the access to customer data (social security numbers, account numbers, etc.) only to authorized personnel. In addition, these financial institutions can use a basic usage audit and limit the distribution (for example, e-mail, Web-mail) of the data to internal recipients within the organization, blocking attempts to send it outside. ▲ ‘Pareto’—covers important proprietary and confidential information and business processes with medium or low change frequency. Implementation of the applicable conditional access and usage control rules can be resource-consuming. You can reduce these costs by using content-aware technologies instead of building customized security controls per application screen. ▲ ‘Progressive’—comprehensive, ongoing review of all major information assets and business processes with update milestones. Most companies will not reach that stage as it incurs considerable costs. Nevertheless, this stage is of high importance for government agencies and commercial organizations with extremely valuable proprietary information assets. Conclusion Information assets are the fundamentals of business today. Their secure management is crucial not only for regulatory compliance, but also for a much tighter control over business processes. In the future we will likely see increased effort and progress in that course, driven either by legislatures or business. The 12 steps in this article supply a structural, discrete and measurable process to achieve this goal. ¡ Ariel Peled is the CTO of Vidius, Inc. and the president of the ISSA Israeli Chapter. He holds a BSc degree in computer engineering from the Technion, Israel. While at Microsoft, he had a prominent role in the development of one of Windows NT’s subsystems and later on, lead the development of several military projects. Dr. Troyansky currently serves as the chief scientist of Vidius, Inc. He holds a Ph.D. degree in computational neuroscience from the Hebrew University and is involved in computer security research since 2000. He also led the development of several military projects. ©2004 Technical Enterprises, Inc. Reproduction of this document without permission is prohibited. THE ISSA JOURNAL ◆ March 2004