Habitats: Toward An Architecture for the Global Information Grid

John Salasin, Defense Advanced Research Projects Agency (DARPA), Assad Moini, Software Productivity Consortium, and Gwen Williams, Schafer Corporation

The military's vision of future warfare as described by Joint Vision 2010 (JV2010) emphasizes broad use of advanced Information Technologies (IT) to significantly improve traditional military capabilities. It implies tight integration and interoperability among DoD systems both horizontally across different domains (e.g., intelligence, operations, and logistics) and vertically between command levels (e.g., National Command Authority, Commander-in-Chief, Joint Task Force Commander, and Small Unit Operations). This tight integration will also include dynamic multi-national, multi-echelon teams engaged in continuous (joint) collaborative planning and/or engagement.

JV2010 also stresses the importance of information superiority and decision superiority as the key to the US military superiority in the 21st century. Decision superiority is the state of having the ability to collect, fuse, process and disseminate an uninterrupted flow of information and command, while denying an adversary the ability to do the same. Of course, information superiority is not an end to itself but a means to achieving decision superiority. The objective is to leverage and transform information superiority into a state where better decisions are made. The overwhelming theme in all of this is the enabling of both greater information sharing and effective, fluid collaboration among heterogeneous units, organizations and systems as dictated by the evolving mission needs. Achieving and maintaining information superiority places a huge emphasis on operational application of information technology. As a result, our success in future operations will heavily depend on our ability to rapidly acquire, assimilate and leverage emerging information technologies.

To address the JV2010 needs, several DoD programs are currently underway including the Navy's Network-Centric Warfare (NCW), the Air Force Joint Battlespace Infosphere (JBI), and the Army's Future Combat Systems (FCS). All these programs share and rely on a common construct called the Global Information Grid (GIG), an Internet-like network of networks, providing the necessary substrate for connecting sensors, shooters, commanders, analysts, legacy and future command and control (C2) and warfighting assets. The GIG serves two main functions. First, it is a notional construct (a conceptual model) for presenting, evaluating, calibrating and unifying advanced military research and technology planning required to realize the JV2010 vision. Second, it is the physical infrastructure that will realize the goal of migrating the DoD warfighting enterprise toward a globally integrated command and control platform, supporting a broad spectrum of future warfighting needs, including operations other than war. Much of the previous work on the GIG has focused on connectivity, networking and basic interoperability issues, such as how to incorporate airborne nodes into the GIG and integration of newer, IP-based networking backbone and the legacy, military radio communication protocols. Among the areas yet to be addressed, are the issues revolving around the control, allocation, monitoring and management of the GIG resources.

We are proposing habitats as a unifying systems concept and an architectural construct for identifying, formulating, assessing and presenting issues, concepts, engineering principles, methods and solution approaches required to make the Global Information Grid a reality. The habitats concept also serves as a test bed for ideas and a focal point for integration of a broad range of research issues into a cohesive research agenda.

The next two sections briefly characterize the GIG environment and the key issue of control and management of the GIG resources, respectively.

The GIG has been characterized as an information environment providing value-added services supporting uninterrupted, secure flow of mission-critical planning and survival information in a globally disbursed environment among producers and consumers. In this respect, the GIG is more than just a network of networks; it is a system of systems consisting of both physical and logical components. As an architectural construct, the GIG promotes a service-centric approach as opposed to the earlier DoD application centric platforms.

To be effective, the GIG must provide the following capabilities:

• Serve as a generic, plug-and-play infrastructure, accommodating rapid, ad-hoc adaptation and reconfigurability to support evolving mission needs, for a broad spectrum of operations including joint, coalition as well as operations other than war

• Provide a seamless environment for end-to-end access to, fusion, dissemination, and presentation of information services regardless of the providers' underlying platforms or networks

• Provide the required computing resources (connectivity, bandwidth, processing) on demand as well as on an anticipatory basis

• Provide management and control mechanisms for provisioning both local and global resources

• Allow for easy extensibility with respect to infusion and insertion of new technologies, while maintaining backward compatibility with the legacy and heritage systems as well as the existing suite of military-unique networking and communication protocols

Implementing the GIG requires dealing with numerous technical challenges especially with regard to end-to-end interoperability and security, as well as organizational issues related to the ownership, control and management of assets and resources linked by the GIG. From a purely technical standpoint, building a global information infrastructure without capitalizing on currently available commercial infrastructure, such as the commercial Internet and Web technologies, is not feasible. In fact, adopting and leveraging these existing commercial technologies in building the GIG is a financial and time-to-market imperative. However, because commercial technologies are designed with a different set of priorities, a number of issues must be addressed.

During this decade, Moore's Law projects a 32-fold increase in processing power and a 256-fold increase in storage capacity; all at the same cost as today's leading edge commercial technology. During the same period, advances in optical networking and wireless technology are expected to produce a dramatic increase in networking bandwidth and connectivity at considerably affordable prices. These technological advancements raise two issues. First, the number of entities situated on and linked by the GIG can potentially be in the billions, making it nearly impossible to monitor, control or manage as a single domain. Second, much of this enabling technology is and will be proliferated throughout the world, allowing potential adversaries to achieve a greater level of sophistication in their military capabilities. A key challenge for the GIG architecture is the ability to rapidly accommodate, assimilate and co-evolve with the emerging commercial technologies.

The GIG must support uninterrupted flow of mission-critical planning and survival information. The highly decentralized nature of the GIG raises serious concerns with respect to the timely and predictable allocation and recruitment of geographically dispersed resources in a manner that is consistent with evolving operational needs and priorities.

The ability to dynamically discover, authenticate and allocate GIG resources will require new technologies. Decentralized control and management of heterogeneous resources, owned or controlled by multiple authorities, over terrestrial, wireless, radio and satellite networks raises serious issues with respect to authorization, mutual authentication, and coordination for access. Additionally, each local authority may have its own internal policies, checks and balances, which are implemented as a complex hierarchical decision-making process. This makes both finding and securing the required resources even more difficult. Development of authority and accountability technologies would facilitate acquisition and authentication of resources.

Historically, military systems (e.g., command and control) have been built as a collection of disparate, unique platforms with fixed functionality, intended to address specific operational needs. Many of the existing DoD communication systems are vertically integrated to satisfy specific warfighting requirements. Many legacy systems have built-in, military unique communication equipments. Achieving interoperability and information sharing among these systems, even at a limited level is not simple.

Habitats are computational building blocks for arranging and managing dynamic relationships in rapidly changing, decentralized environments, such as the GIG. Habitats provide a computational model for organizing, coordinating and managing (human-to-human, human-to-systems or system-to-system) interactions in environments where independent actors must heavily rely on information and communication infrastructures to arrange and coordinate interdependent (joint) activities. These activities may typically cross multiple organizational and functional boundaries. As such, habitats provide the necessary shared context for integrating business processes across independent organizations.

Conceptually, habitats resemble virtual (human) organizations, dynamically formed to augment or extend human organizations in support of a particular business activity or mission. Therefore, habitats can be either mission-centric or organization-centric. They serve to provide a shared context and transactional boundary for ongoing activities involving humans, agents and systems. Habitats also serve as shared collaboration spaces mediating secure, reliable and transparent access to remote C2 and warfighting assets, irrespective of their (network) location or underlying platform, in a manner consistent withorganizational policies and procedures and subject to appropriate checks and balances.

Habitats exploit the power of networks to bring together geographically dispersed, distributed entities (human actors, warfighting assets, C2 infrastructure) toward accomplishing a specific task or mission. As such, they go beyond merely facilitating information sharing; they serve as a conduit, arranging new forms of interaction and collaborative relationships.

The goal of habitats science and technology is to provide the necessary technical infrastructure needed to ensure that independent groups or organizations, from distinct spheres of powers, can effectively and expeditiously align their activities to achieve a common objective. Toward this end, the habitats technology must address organizational, operational and technical issues that arise in the context of ad-hoc coordination, collaboration and interoperability among otherwise independent organizations and systems.

Habitats also serve to provide a computational context and mechanism for partitioning and scoping activities and automated (business) processes that span multiple systems and organizational boundaries. As such, they provide transactional boundaries necessary for compartmentalizing, monitoring, auditing, or constraining activities that may span heterogamous platforms and networks. They also provide a unified computing model that combine and subsumes the traditional distributed objects and hypermedia web document model.

The following section illustrates the notion of habitats in the context of a combat situation, using an operational scenario. Today, much of the required collaboration is complicated and hampered by the fact that many of the functional units and systems are often spatially dispersed and have hard-wired communication line (stove piped). The concept of habitats is used to dynamically organize otherwise stand-alone systems into a collaborative, networked system of systems.

In this section we present a set of desiderata (necessary characteristics) that must be fulfilled by the future habitats technology. To meet the following desiderata, the habitats technology must provide universal (platform and operating-system independent) supporting mechanisms at the GIG infrastructure level.

• Dynamic Membership Model. New members are admitted or enlisted and old ones can leave
• the habitats (voluntarily or become disconnected) or their membership can be revoked if their integrity is compromised. Traditional networks generally rely on the existence of a permanent, centrally managed security service. In habitats, the entire functionality associated with security is fully distributed across all the members. This is necessary to ensure a high degree of resilience to unexpected (communication) failures or removal of a member. The key elements of the habitats dynamic membership model are: the mutual authentication of habitat members, proof of habitat membership to outsiders, and authentication of members with special privileges. The dynamic membership model is required for incorporating or removing entities (e.g., coalition partners, non-DoD, civilian agencies, or allies). Habitats must support flexible and interoperable policy management and enforcement by providing a dynamic membership model. Membership in secure habitats may be required to facilitate total resource management as well as access to data.
• Dynamic, Granular Trust Model. Military organizations often need to cooperate or interoperate with civilian or non-governmental organizations or members of military alliances toward a common goal. In many cases, it may be necessary to share or transfer highly sensitive information, for example by providing access to intelligence or reconnaissance assets on a temporary basis subject to limited, asymmetric trust. Additionally, given the evolving nature of operations and alliances, the level and granularity of trust among the partners may change rapidly. Habitats must explicitly support the notion of granular trust relationships (coarse and fine grain trust levels) as well as an asymmetric trust model. Habitats infrastructure must provide the necessary mechanism to ensure end-to-end authority accountability across interacting, overlapping habitats.
• Dynamic Policy Management. Coordination policies highly depend on the context of the activities and the nature of collaboration. Even within the same context, policies often may have to evolve due to the changing nature of the collaboration itself (with respect to both scope and extent). In habitat relationships, for example in a coalition or operations other than war (OOTW) setting, the level of trust among participants (humans, military and non-governmental organizations) can change over time, thus necessitating co-synchronization or updating policies within and across interacting habitats.
• Boundlessly Scalable. Scalability is the ability to increase or decrease the size and/or scope, as necessary, to accommodate dynamic changes in operations, caused by the changes in the environment. Habitats will be capable of expanding (upsizing) and retracting (downsizing) bandwidth and internal processing capacity relative to size, scope and quality of service. Scaling up is the ability to dynamically expand or extend operations as the workload grows. Scaling down is the ability to contract operations in size or scope, for example by shedding functionality or resources, in response to diminishing demands, under consumption of resources or simply to cope with independent local or remote failures. Expandability and adaptability are two complementary aspects of scalability.
• Expandable. Expandability refers to a change in size or scope necessitating a change in structure or organization of habitats. Expandability in habitats requires the ability to create new habitats practically anywhere, anytime. New habitats can be bootstrapped from existing ones, for example, by federating existing ones. At the same time, existing habitats must be able to incorporate new habitats within themselves and be able to create new (sub-) habitats by reorganizing, combining or regrouping.
• Adaptable. Adaptability is the ability to evolve habitats structure (roles, responsibilities, authority and accountability relationships) in response to the evolving mission needs caused by external factors, while meeting commitments in place (e.g., carrying out activities already in progress). Adaptability demands reorganization by amending or extending existing habitats by incorporating new entities and/or assets.
• Transitory in Nature. Habitats are dynamic. Once formed, they continue to co-evolve with changing mission needs and objectives. They are ultimately dissolved or placed in a state of hibernation; habitats placed in hibernation can be resurrected and placed back into the operational mode as the need arises. In this respect, habitats are organizational/operational blueprints for rapid creation of electronic organizations (rapid strike teams/cells) required to support specific mission needs.
• Overlapping. Habitats can overlap in space and time. Two habitats can overlap in space by virtue of sharing common entities or assets (e.g., the same ground sensor or airborne node can be concurrently assigned to multiple habitats, subject to different access and usage rules). Two habitats can temporally overlap because of their mutual dependence on completion status of events or activities hosted by another habitat.
• Managing Quality Of Service. Habitats infrastructure must provide required computing resources (connectivity, bandwidth, processing) on a just in-time, anticipatory, or reservation basis.
• Unified Control and Management. Habitats infrastructure must provide universal (platform and network independent) mechanisms for management and control of both local and global resources.

Consider a Time Critical Target (TCT) cell habitat. The function of the TCT cell is to assign resources to emerging high priority ground targets, identified by Joint-STARS (JSTARS). This habitat is instantiated (formed) when a Moving Target Indicator (MTI) track appears and JSTARS produces a track report. The Air Defense System Integrator System (ADSI) automatically assigns a System Track Number (STN) and begins reporting the track with updates to attributes such as location, speed, direction, and classification (e.g., friendly, enemy or undetermined).

The ADSI then evaluates the track and correlates JSTARS information with that from other intelligence sources, e.g. point of origin or communications intelligence (COMINT) reports on payload.

The TCT habitat is dynamic. Based on its location (e.g., the theatre of operation) it has the ability to interoperate with certain specific Information, Surveillance and Reconnaissance (ISR), surface-to-surface, Combat Operations, or Area Air Defense (assets) habitats. As a background task, it can update its awareness of and access to information about resources in its current area of concern (based on the JSTARS location). It is activated on nomination of a dynamic target and hibernates when no targets are active (i.e., the Dynamic Target Queue is empty).

The TCT cell habitat provides specialized services to its components. These might include various models (e.g., dealing with munitions effectiveness, platform/resource performance, meteorological effects, and available bandwidth/latency). It provides access to specialized, context-dependent data (e.g., enemy order of battle). It coordinates components (e.g., resource location and status agents, time sensitive resource data update tools, tools for coordination and deconfliction). It also enforces various rules or constraints (e.g., conditions requiring different levels of authorization, allowable threshold of potential collateral damage, and required authorization), commits authority levels, required sequence of operations (process), rules for lawful targets (and when they need to be enforced), fault-tolerant fallbacks (QoS, data availability concerns), and access and update of authorization capabilities and permissions. All of these models, services, and rules are context sensitive ­ they may depend on the theatre involved or the state of hostilities.

Note that this scenario and its follow-on activities require dynamic recruitment of resources and services ­ a primary responsibility of the habitat. These resources may include, for example, information about potential second sensor sources and appropriate/available weaponry assets. This information might come from other (than Air Force) services/organizations. It could include: sensor or weapon capabilities, current tasking/scheduled tasking, current location, time to image or reach target, tasking authority, time sensitivity of data, and weather.

The Table 1. illustrates the current situation and the manner in which habitats will address specific present needs and shortcomings.

Today, the U.S. military is in the process of transforming itself into a nimble, cohesive organization of the information age by replacing the rigid, stove piped warfighting infrastructure of the past with a more flexible, agile organizational structure equipped with distributed sensing, planning and execution capabilities, all networked together via the Global Information Grid. In this networked environment, the relationships and the resulting interactions are substantially different from the traditional hierarchical command and control model. To successfully operate in this environment, the traditional top-down C2 hierarchy must be modified to allow a more decentralized, peer-to-peer decision process at the lower levels, across units, echelons, organizational and functional boundaries. The habitats concept and its supporting technology will enable a new generation of systems that can successfully and predictably operate in a network-centric world of vastly distributed and dispersed resources, devices and users interacting via the GIG.

---