Autonomic computing systems arose from the notion that complex computing systems should have properties like those of the autonomic nervous system, which coordinates bodily functions and allows attention to be directed to more pressing needs. An autonomic system allows the system administrator to specify high-level policies, which the system maintains without administrator assistance. Policy enforcement can be done with a rule based system such as Jess (a java expert system shell). An autonomic system must be able to monitor itself, and this is often a limiting factor. We are developing an automatic system that has a policy engine and uses Artificial Immune Systems (AISs) to sense its environment and to monitor its components and performance. AISs emulate the natural immune system to defend the body against external malicious entities. The proposed system monitors itself without human intervention and thus addresses the problem of systems complexity.
With more systems and devices networked together, more system administrators are required to monitor and maintain these networks, and maintenance becomes costly and time consuming. Yet it may not be clear how a complex system may manage itself in the absence of a human. We here report on an autonomic system we are developing for a network-connected computer system that self-manages by using mechanisms similar to the autonomic nervous system of the body to achieve the so-called self-CHOP properties: self-configuring, self-healing, self-optimizing, and self-protecting. A rule-based policy enforcement engine is used (like the body’s hypothalamus) to regulate the autonomic system, and the system has artificial immune systems (AIS) that mimic the functions of the sensory and motor subsystems.
The remaining sections of this paper are organized as follows. The next section provides background on autonomic computing, AISs, and rule-based systems. We then discuss the motivation for implementing an AIS. There follows a section that presents the architecture of our autonomic system. This is followed by the AIS sensoring agent section, where we describe how the AIS aids the autonomic system with its self-CHOP capabilities via anomaly detection. The context monitor is then described, followed by a section describing how the self-CHOP properties are realized by our system. The final section presents the conclusion and future work.
IBM, in a 2001 manifesto, compared complex computing
systems to the human body, which has an autonomic nervous
system that removes the tasks of consciously coordinating
bodily functions
IBM also articulated the four self-CHOP properties.
Selfconfiguration is defined as components and systems being
configured as per high-level policies
The natural immune system is a defense mechanism that can
learn about foreign entities that enter the body and respond
to them by creating defensive antibodies. This concept has
been artificially simulated for intrusion detection, resulting
in an approach known as an artificial immune system (AIS)
AISs have also been applied to the problems of fault
diagnostics, fraud detection and detecting viruses
The classical application of rule-based systems is in expert
systems, which typically use a human experts knowledge for
solving real world problems
Carey et al. created a composite service execution engine
for composing web services
We aim for a system capable of adapting to unforeseen occurrences in the operating environment. This includes adjusting to situations in a proactive manner as well as reactively supporting system recovery. Here pro-action amounts to perceiving danger then preempting harm or any compromised system state. Fault tolerance in the operating environment permits system execution to continue without interruption, while preempting failure from a system-wide perspective presupposes the occurrence of a compromised system component or operation.
Anomaly detection can be a key component to
properly assessing whether events are uncharacteristic of a
system’s configuration
One immunity-inspired algorithm of interest focuses on
danger theory, which extends the role of the innate
immune system for discriminating between ‘self’ and
‘nonself’ but employs the acquired immune system to react to
danger
We here discuss how the AISs contribute to the self-CHOP properties. Note that sometimes we refer to a single AIS while other times we partition the functionality among several AISs. Although we refer to sensing, note that the AISs detect both external and internal state. The AIS sensoring apparatus will consist of detector populations that monitor system activity as does a network sniffer or intercepting proxy. The AIS component will deploy detectors that compare system activity against normal system behavior. The detection methods behave as intrusion detection systems (IDSs) that detect system irregularities not stated in the innate system configuration. The detection method can also adapt to changes in the system.
The detector set will become a new feature vector for detecting danger in the system. The new vector will allow the system to recognize system behavior that has been previously labeled as dangerous. The new danger vector is then propagated throughout the system to scan for stress levels that are approaching this danger state.
Each round, the AIS will acquire more knowledge about the operating environment via the detector populations. The cloning process allows one generation to be more adept than the previous. The acquired immunity that aids the AIS in determining the health of the system is expressed via the number of danger vectors that have been detected and mitigated properly, with more emphasis on the mitigation process.
Following the detection of a stress level threshold being met, the policy engine will determine whether the system is in danger based upon the signal that the AIS found to indicate danger. If danger is confirmed, then the policy engine will determine what the system needs to repair or heal itself. For each consecutive detection round, the AIS will monitor the healing process by executing a subset of detectors for repair monitoring.
The system will adapt to the configuration expected of
systems and components as per high-level policies during
self-configuration
The environment of a computer system can undergo changes
for any number of reasons. To deal with this, one
exploits context awareness, a computer sensing and reacting
to changes in its environment
For our rule-based policy engine, we used Jess (Java
Expert System Shell)
One of the purposes of our proposed system is intrusion detection. With this in mind, there are policies that are focused on this activity. For one thing, if packets from the network come in and the timestamp of these packets are a certain time away from the system’s current time, a flag will be raised. If continuous traffic is coming from one specific IP address, it could be indicative of a Denial-of-Service attack. In this case, the system may block packets from that particular IP address. For any input coming into the system, it is assumed that the input is in a standard format. However, if the input appears to be irregular, such as a password having many symbols not numerical or alphabetical, then a warning flag may be raised. There are also policies related to internally monitoring the system. The system should run fairly consistently, where the CPU usage may spike or idle depending on certain actions occurring on the system. If there are times when the CPU spikes or peaks and none of the activities that normally causes this are occurring, then the system may take actions to run a diagnostics check on it. If enough memory is consumed, the system may run a heuristic to delete items that it considers to no longer be necessary. The AISs in the machine are developing detectors to detect anomalies externally and internally. The policies can have additional steps that state that if it catches anything the detectors from the AISs do not, it will prompt the AISs to adjust its detector creation strategy.
enough memory is consumed, the system may run a heuristic to delete items that it considers to no longer be necessary. The AISs in the machine are developing detectors to detect anomalies externally and internally. The policies can have additional steps that state that if it catches anything the detectors from the AISs do not, it will prompt the AISs to adjust its detector creation strategy.
Within any automated system, faults can occur. Faults can
be described as unexpected changes from the normal system
condition. The area of research dedicated to detecting faults
is referred to as the Abnormal Event Management (AEM)
in research done by Laurentys et al.
A major aspect of protection is sensing attempted intrusion, which is a standard task for AISs, detecting “self” and “nonself.” We can generalize this function to perceiving danger (in the environment) in general and perhaps even to recognizing a need to adapt to a change in the environment. The general notion is that of monitoring the environment to protect self. The intent is that an AIS will identify a threat and characterize it sufficiently so that the policy engine may activate resources as per the applicable policies. The expert system may consult with AISs in the course of enforcing a policy.
In a similar vein, we can have the AISs detect when the system is not “itself” and in need of repair; this is the heal CHOP attribute, and what is monitored is self. The role of the AIS here is to identify faulty behavior and to characterize it so that the expert system may activate resources as per the applicable policies. Again, the policy engine may consult with AISs in the course of enforcing a policy.
This vigilance regarding the system’s own behavior may extend to self-optimization if we have a baseline reading of system behavior. An AIS would help us distinguish acceptable patterns that deviate from this pattern (as when some resource is accessed) from unacceptable patterns. Something similar could address self-configuration. The most obvious occasions for self-configuration, however, are where a new device is attached. In such cases, a simple signal from a context monitor would suffice to provide the policy engine all the information needed to apply policies.
We have sketched the autonomic system we are developing that uses artificial immune systems (AISs) augmented with a context monitor to provide data to a rule-based policy engine. The architecture is conceptualized as the AISs and context monitor providing sense data to the policy engine, but not that the data is also from the internal state of the system. We discussed how the system supports the self-CHOP properties: self configuring, self-healing, self-optimizing, and self-protecting.
We are implementing our autonomic architecture on a stock workstation that is attached to the Internet (inviting intruders) and to which we can attached multiple accessories (requiring self-configuration). In the future, we plan to investigate autonomic cyberphysical systems. In the future, we shall look into other biological metaphors and implement some into our system.