Mohammad Rabby

Software cybernetics to infuse adaptation intelligence in networked systems

Future network systems are expected to have various levels of adaptation capabilities: at parametric level, service-level, and application-level. These capabilities are often realized in multiple system layers, with the control logic needed for a specific capability residing in the application agents interfacing with the underlying network system services. We employ a software cybernetics approach, where an intelligent physical system module (IPW) embodies the core adaptation functionality to respond to the changing environment conditions and user inputs. The IPW exhibits an intelligent behavior over a limited operating region of the system. And, it is augmented by a supervisory module (ICW) that houses external feedback loops for situational assessment of the changing environment conditions. Our autonomic management of various hierarchical control loops comes under the ambit of Cyber-Physical Systems (CPS). The ICW patches the IPW with suitable control parameters and rules/procedures when the system operating conditions change. The paper illuminates our concept of IPW with a software engineering-oriented case study of multi-source video transfer over a bandwidth-limited network path.

Distributed adaptation algorithms for rate-controlled video multicast over shared infrastructure networks

We consider a wide-area video conferencing application where the video sources can adapt their send rates according to the available bandwidth in the network paths. We advocate a joint rate control of the sources to relieve the congestion, instead of running multiple instances of a single-source adaptation algorithm and additively superposing their results. The existing techniques work nicely with single-source trees, but do not work optimally in the case of multi-source trees with shared QoS goals. Using the well-known AIMD-based adaptation procedures, we incorporate the topology inferencing mechanism into a coordinated rate adaptation algorithm executed by the loss-experiencing sources. The paper provides a simulation based evaluation of our algorithm to corroborate the benefits.

Performance engineering of replica voting protocols for high assurance data collection systems

Real-time data collection in a distributed embedded system requires dealing with failures such as data corruptions by malicious devices and arbitrary message delays in the network. Replication of data collection devices is employed to deal with such failures, with voting among the replica devices to move a correct data to the end-user. Here, the data being voted upon can be large-sized and/or take long time to be compiled (such as images in a terrain surveillance system and transaction histories in an intrusion detection system). The goal of our paper is to engineer the voting protocols to achieve good performance while meeting the reliability requirements of data delivery in a high assurance setting. The performance metrics are the data transfer efficiency (DTE) and the time-to-complete a data delivery (TTC). DTE captures the network bandwidth wasted and/or the energy drain in wireless-connected devices; whereas, TTC depicts the degradation in user-level QoS due to delayed and/or missed data deliveries. So, improving both DTE and TTC is a goal of our performance engineering exercise. Our protocol-level optimizations focus on reducing: i) the movement of user-level data between voters, ii) the number of voting actions/messages generated, and iii) the latency caused by the voting itself. The paper describes these optimizations, along with the experimental results from a prototype voting system.

Autonomic Management of Replica Voting based Data Collection Systems in Malicious Environments

We describe a model-based approach to QoS management in a replica voting based data collection system. The voting among replicated data collection devices achieves trusted data delivery to the end-user in a hostile environment: such as data corruptions by malicious devices and security & bandwidth attacks on the wireless data paths. How often an accurate data is delivered to the user in a timely manner depicts the QoS of data collection system. Aided by a computational model of the voting system, a situational assessment module macroscopically controls the voting system core based on the sensed external events. Our goal is the optimal use of system resources while enforcing an acceptable QoS. The paper describes the management methods for autonomic control of the degree of device replication and/or the algorithmic parameters (e.g., wireless bandwidth allocation) in response to the dynamically changing QoS needs and environment conditions. Our management methods are reusable across different systems, which lowers the software costs in the development of such complex systems.

Cyber-physical systems based modeling of adaptation intelligence in network systems

The paper uses the cyber-physical systems (CPS) framework to model the intelligent adaptation behaviors in complex network systems. The CPS framework is anchored on intelligent physical worlds (IPW) around which complex adaptation behaviors are built. An IPW is an embodiment of control software functions wrapped around the raw physical processes (such as routers, links, hosts, and protocols), performing the core communication activities while adapting its behavior to the changing network conditions and user inputs. The IPW exhibits an intelligent behavior over a limited operating region of the network system, which is in contrast with the traditional models where the physical world is basically dumb. To perform over a wider range of network operating conditions, the IPW interacts with an intelligent computational world (ICW) to patch itself with suitable control parameters and rules & procedures relevant in those changed conditions. The modular decomposition of a network application into IPW and ICW lowers the design complexity of network systems, and simplifies the system verification/testing. The paper illuminates our CPS-based approach with a case study of adaptive video transport over a bandwidth-limited network.

Bandwidth measurement and management for end-to-end connectivity over IP networks

The paper describes a policy-based model for cost-effective ‘data connectivity’ provisioning between session-level end-points. The connectivity provider (SP) may employ an architecture for end-to-end QoS control between data aggregation points. It involves: i) maintaining multiple diffserv-type connections between end-points with parameterizable QoS differentiation between them, and ii) admission control at end-points with intserv-type bandwidth management over connections. (ii) aggregates data flows with closely-similar QoS needs over a single end-to-end connection. (i) apportions the available infrastructure bandwidth between various end-to-end connections that carry (aggregated) data flows with distinct QoS levels. Flow aggregation over a connection allows reaping the statistical multiplexing gains in bandwidth, i.e., meets the SPs revenue incentives. Whereas, connection-level bandwidth allocation allows meeting the QoS needs of data flows, i.e., guarantees the end-users utility. The management functions of SP monitor the changes and/or outages in network bandwidth in a dynamic setting (as in IP-based networks), and maps them onto the connectivity costs incurred for QoS control. Our model allows installing policy functions at end-points for cost-optimal connectivity provisioning.

Protocol-level reconfigurations for autonomic management of distributed network services

The paper describes a service model for the autonomic management of distributed networked systems. In this model, a service provider (SP) maintains multiple protocol modules to exercise the infrastructure resources under various environment conditions. Each protocol exhibits a certain degree of performance optimality and service resilience in distinct operating regions of the network infrastructure and the environment. During run-time, the SP selects one of the protocol modules that can meet the application-requested Quality of Service (QoS) obligation against the prevailing operating conditions. Under normal conditions when the external disturbances are benign (e.g., low packet loss in the network), an optimal usage of the network resources is important. Under adverse conditions however (such as prolonged sub-system outages and failures), a sustained access to the network service at some minimum acceptable level becomes more important than a resource-optimal service offering. Often, a resilient protocol incurs more resource usage to tackle the hostile environment conditions than a performance-conscious protocol tuned for the normal case operations (‘a single shoe does not all sizes’ !!). Accordingly, a protocol selection by the SP considers the trade-off between ‘service availability’ under extreme operating conditions and ‘resource optimality’ under normal operations. Our model allows a ‘dynamic switching’ from one protocol module to another at run-time based on the changing environment conditions. The paper advocates ‘protocol switching’ as a foundation for building autonomic network services with the twin goals of service-level availability and performance.

QoE-cognizant video rate adaptation over bandwidth-limited network paths

We employ a model-based engineering (MBE) of video rate adaptation functions on a bandwidth-constrained end-to-end network path: such as the Internet. The intrinsic complexity of video rate adaptive system arises from two estimation problems: first, the available bandwidth along the path, and second, the bandwidth demand of a video data flow exercised on the path. The well-known AIMD (additive increase multiplicative decrease) algorithm works reasonably well to get around these difficulties, albeit, with certain limitations that affect the QoE (quality of end-user user experience). The AIMD algorithm is unable to distinguish between the scenarios of a congestion in the network and a marginal over-rate allocation, as it searches for an optimal send rate of video. Consequently, AIMD does not have good handle on the convergence latency during a congestion relief, and the rate jitter incurred in the steady-state. In this paper, we provide extensions to AIMD in the form of dynamic plugin of the rate adjustment parameters, namely, how fast or slow the rate should be reduced or increased. The extended AIMD, armed with a system-level support mechanism for dynamc parameter adjustments, improves the overall QoE of end-users when reacting to congestion. The paper describes experimental results to demonstrate the viability of our extended AIMD.

Situational context for replica voting based data collection in hostile environments

The paper discusses the infusion of context awareness in the management of data quality in a replicated sensor based data collection network. Given a 2-phase voting protocol to decide on data delivery from sensors to the end-user, the paper discusses the impact of environment parameters (such as fault severity and network message loss) and resource parameters (such as network bandwidth and number of replicas) on the quality of user-level data delivery. The degree of control is determined by the contextual inputs that capture the dynamically changing scope and importance of data to the application.

Management intelligence for optimal resource allocations in network server systems

In this paper, we provide a control-theoretic treatment of the resource allocations that adaptively occur in a QoS-aware network server system. Here, the target system being controlled is a logical service point that processes the transactions requested by clients using a resource infrastructure, with a goal of maximizing the revenues. Accurate management of resource allocations with a revenue-oriented goal is quite complex, due to the interactions among various transactions that dynamically share the resources in the system (such as server nodes, disks, content caches, and network bandwidth). So, we adopt an on-line monitor-and-control approach, aided by heuristics, that iteratively adjusts the resource allocation based on the observed transaction drop rate. We undertake a case study of end-to-end QoS-adaptive data transfer to illustrate the methodology. In terms of control theory, the bandwidth allocation and the packet loss rate constitute the system input and output respectively, with the heuristics-based bandwidth adjustment strategies incorporated in a controller along the feedback loop. The use of control theory allows offering predictable convergence properties of the QoS seen by applications, while maximizing the service provider revenues.