K. Appleby, G. Goldszmidt, and M. Steinder. Yemanja -- a layered event correlation engine for multi-domain server farms. In IFIP/IEEE International Symposium on Integrated Network Management VII, Seattle, WA,  Home/Search   Document Details and Download   Summary   Related Articles   Check

....between abstract functions and services between neighboring layers to be modeled. It does not, however, provide means to represent multiple possible failure types of a particular function or service, nor does it provide any support for modeling non deterministic system behavior. Yemanja [1] defines a hierarchical set of entity models, which correspond to physical or conceptual network entities, e.g. network layers. Entity models are instantiated with the data obtained from the arriving event attributes or from the configuration database, and communicate using internal composite ....

....internal composite events. As events propagate up the entity model hierarchy, their semantics becomes more and more abstract so that higher level scenarios can correlate higher level symptoms with lower level failures without the detailed knowledge of the low level network state. Both Yemanja [1] and the layered system model [5] address the issue of vertical fault propagation. They do not discuss the situation in which an explanation of a higher level problem (e.g. end to end service failure) must be selected from a large set of possible lower level failures (i.e. host to host link ....

K. Appleby, G. Goldszmidt, and M. Steinder. Yemanja -- a layered event correlation engine for multi-domain server farms. In IFIP/IEEE International Symposium on Integrated Network Management VII, Seattle, WA,

....across multiple protocol layers, because different time windows may apply to different layers, and (3) it is inefficient because the entire computational effort is performed at the end of a time window instead of being distributed over time. It is believed that event driven fault localization [2, 12] is a better approach to the problem. Preferably, the final solution should be created incrementally based on the solutions available after previous symptom observations. Real time fault diagnosis Substantial loss of revenue may result if faults within the enterprise communication system are not ....

....this goal requires fault localization techniques with low computational complexity. 5 A NEED FOR END TO END SERVICE FAULT Recent publications on fault localization recognize the need for multi layer fault localization and try to address some of the issues described in Section 4. Yemanja [2], a model based reasoning system, tries to overcome the difficulties involved in managing system models by defining a set of entity models, which correspond to physical or conceptual network entities, e.g. network layers. It avoids maintaining an explicit network model by providing scenario ....

[Article contains additional citation context not shown here]

K. Appleby, G. Goldszmidt, and M. Steinder. Yemanja -- a layered event correlation engine for multi-domain server farms. In IM'01 [16].

.... services to end systems, a management system needs to efficiently and accurately identify the occurring network failures [13, 25] A common procedure is to correlate network or service layer symptoms; however, this process is usually impaired by the large number of a system s layers and parameters [1, 8, 22], their interactions, and the uncertainty about their state. This paper presents a preliminary study of applying statistical techniques [16, 18] known in the engineering quality control to cope with the exponential complexity that often hampers the event correlation process. The concept of ....

....services. The problems typically addressed in the literature are (1) incomplete knowledge about the existence of causal relationships between network events [13] 2) possibility of incomplete symptom observations or spurious symptoms [9, 25] 3) system adaptability to configuration changes [1], 4) ability to learn event correlation patterns [14] and (5) temporal event correlation [15] Most of these techniques rely on the assumption that the existence of multiple simultaneous faults is negligible. Relationships between network objects are often represented using a dependency graph ....

[Article contains additional citation context not shown here]

K. Appleby, G. Goldszmidt, and M. Steinder. Yemanja--- a layered event correlation engine for multi-domain server farms. In IFIP/IEEE Intg'ed Network Management VII, Seattle, WA, 2001.

....resource availability related problems such as a broken cable or an inactive interface. Recently, the scope of fault localization has been expanded to include the diagnosis of performance problems in higher layers of the communication system, such as the transport and application layers [1] [2]. For this purpose, non deterministic reasoning needs to be incorporated in the fault localization process [3] This paper utilizes belief networks [4] to perform fault localization in communication systems whose failure propagation model may be described by bipartite causality or dependency ....

....(such as a broken cable or an inactive interface) since their impact on other network components is easily predictable. Recently, more attention has been devoted to diagnosing performance related failures in higher layers of the protocol stack including the transport and application layers [1] [2], 3] In this application, the deterministic model is not sufficient because the impact of performance related failures in one entity on dependent entities is not easy to predict. Also, the dependencies among system entities change so frequently that they may not be deterministically captured by ....

K. Appleby, G. Goldszmidt, and M. Steinder, "Yemanja -- a layered event correlation engine for multi-domain server farms," in Integrated Network Management VII, G. Pavlou, N. Anerousis, and A. Liotta, Eds. May 2001, pp. 329--344, IEEE.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC