Efficient Over-Provisioning of Network Systems and Services: Principles and Practices Dong Xuan Department of Computer Science and Engineering The Ohio-State University http://www.cse.ohio-state.edu/~xuan 2/24/2019 The Ohio State University
What is Over-Provisioning? Resources are allocated conservatively, depending on expected demands Examples: replicated content, replicated servers, allocating more bandwidth, multi-path routing etc. 2/24/2019 The Ohio State University
The Ohio State University Outline Objective Principles Practices in Overlay Networks Practices in Sensor Networks Final Remarks 2/24/2019 The Ohio State University
The Ohio State University Objective Providing high performance, reliability and security to network systems and services 2/24/2019 The Ohio State University
Challenges and Opportunities Traffic amount Dynamics of traffic pattern Malicious and non-conforming participants Opportunities: Resources, such as bandwidth, storage, processing power are no longer the bottlenecks that used to be so in the past 2/24/2019 The Ohio State University
Why Over-Provisioning? Enable uninterrupted services Reaction under extreme operating conditions are milder if not eliminated Maintenance and corresponding dynamics are easier if done properly System update is easier 2/24/2019 The Ohio State University
The Ohio State University However…… Over provisioning is not always good Over provisioning also comes at the price of increased maintenance Resource come at a price, they are not free Resource availability is unbalanced 2/24/2019 The Ohio State University
The Ohio State University What We Want to Do? Study the principles of over provisioning Practices in a wide spectrum of network systems and services 2/24/2019 The Ohio State University
The Ohio State University Related Work Bandwidth over-provisioning by ISPs (Internet Service Providers) Data backup for fault tolerant services Over-deployment in sensor networks 2/24/2019 The Ohio State University
The Ohio State University Principles A case study – bandwidth over provisioning in networks Currently it is conducted in an ad hoc manner by ISPs QOP: Quantitative Over Provisioning Our work on Transaction on Networking 04 [1] and RTSS 01 [2] 2/24/2019 The Ohio State University
Further Study on Over Provisioning Principles System resources System nodes Connectivity Network Paths Data content, energy and storage Dynamics due to failures and attacks 2/24/2019 The Ohio State University
Practical Applications of Over-Provisioning Overlay Networks Sensor Networks 2/24/2019 The Ohio State University
Practices in Overlay Networks Secure Overlay Forwarding Systems Resilient Structured Peer to Peer Systems QoS aware and Reliable Overlay Multicast and Anycast Services 2/24/2019 The Ohio State University
The Ohio State University Overlay Networks 2/24/2019 The Ohio State University
Secure Overlay Forwarding Systems It is an intermediate forwarding overlay system to defend against DDoS attacks Layering: Each node only knows the next layer nodes Access to target controlled by a set of filters Target is known only to filters 2/24/2019 The Ohio State University
The Ohio State University Design Features The number of layers: 3 layers of hierarchy between sources and a target Mapping degree: Number of next layer neighbors Node density: Number of nodes per layer Under random congestion attacks, path availabilities are high if mapping degree is high 2/24/2019 The Ohio State University
The Generalized Secure Overlay Forwarding System We have generalized the system in ICDCS 04 [8] Design features are flexible 2/24/2019 The Ohio State University
Intelligent DDoS Attacks Combination of Congestion-based attacks and break-in based attacks Congestion attacks result in node being non-functional for the duration of the attack Successful break-in attacks result in disclosure of next layer neighbors 2/24/2019 The Ohio State University
System Performance Observation Over Provisioning is not always good Care should be exercised 2/24/2019 The Ohio State University
Resilient Structured P2P Systems Distributed Hash Table (DHT) based Node ID and data ID match together CAN, CHORD, PASTRY and TAPSTRY These systems are not resilient to malicious attacks ! Our solutions: Over provisioning in neighbor connectivity RCHORD [4] and CAN-SW [3] 2/24/2019 The Ohio State University
QoS Aware Overlay Multicast and Anycast Unicast, multicast and anycast Network layer multicast and anycast We have proposed an efficient fault-tolerant multicast routing protocol in TPDS 99 [5] (38) We have proposed a routing protocol for anycast messages in TPDS 00 [6], 04 [7] (38, 39) Overlay multicast and anycast Multiple path over provisioning based approaches 2/24/2019 The Ohio State University
Practices in Sensor Networks Sensor network deployment using limited mobility sensors Defending against Physical Attacks 2/24/2019 The Ohio State University
The Ohio State University Sensor Networks A new paradigm of networking A lot of applications like tracking intruders, monitoring animals, forest fires, and warehouse monitoring Cheap, easy to deploy, but limited in energy Base station A simple sensor network MTS 310 CA sensor 2/24/2019 The Ohio State University
Sensor Networks Deployment using Limited Mobility Sensors Sensor network deployment Issues Sensors may be damaged Sensor may be out of energy Manual redeployment is hard Solutions Over-provision sensor nodes Exploit sensor mobility 1 5 6 3 4 2 7 8 9 10 11 12 13 14 15 16 2D-grid 2/24/2019 The Ohio State University
Limited Mobile Sensors Mobility in sensors is an energy consuming operation XYZ sensor platform can move up to 165 m DARPA has already built limited mobility sensors, whose maximum movement is 100 hops Resource of sensor nodes are redundant but their mobility is limited 2/24/2019 The Ohio State University
Our Deployment Problem Problem definition Given 2-D grid sensor network model, determine a movement plan for the sensors to minimize variance in number of sensors among all regions from and simultaneously minimize the required number of movements Variance = No. of movement hops = 2/24/2019 The Ohio State University
The Ohio State University An Example Sensor Network with 16 regions and =2 A simple, purely localized solution Regions 14, 15 and 16 have less than 2 sensors (b) (a) 1 5 6 3 4 2 7 8 9 10 11 12 13 14 15 16 2/24/2019 The Ohio State University
Discussions on Our Deployment Problem Each region has sensors, which is over-provisioned to provide reliable services It is a non-linear optimal problem. However, when = 1, the problem is changed to a linear one [10] The problem is harder due to over-provisioning 2/24/2019 The Ohio State University
The Ohio State University Our Solutions We proposed two classes of solutions Max-flow based solutions Translate non linear variance problem into linear weight assignment problem Translate sensor network into a graph structure and determine minimum cost maximum weighted flow plan It is optimal if run in a centralized manner Can also execute in a distributed manner Simple Peak-Pit solution Pits request sensors from peaks. Requests contain weights depending on sensors needed Requests are served in descending order of weights Performance is good under favorable deployment conditions 2/24/2019 The Ohio State University
Defending against Physical Attacks in Sensor Networks Physical attacks: destroy sensors physically Physical attacks are inevitable in sensor networks Sensor network applications that operate in hostile environments Volcanic monitoring Battlefield applications Small form factor of sensors Unattended and distributed nature of deployment Different from other types of electronic attacks Can be fatal to sensor networks Simple to launch Defending physical attacks Tampering-resistant packaging helps, but not enough We adopt sensor node over-provisioning approach Physical attacks can permanently destroy the sensors, which are different from electronic attacks such as jamming attacks, which tries to interfere the radio channels and interrupt the sensor networks’ operation. Emphasize that physical attacks are simple to lunch. 2/24/2019 The Ohio State University
Blind Physical Attacks Due to the brute-force destruction methods and blindly selecting attack areas 2/24/2019 The Ohio State University
Search-Based Physical Attacks It is hard for the attacker to get the exact location of the sensors, but it can isolate a relatively small area for each detection sensor. 2/24/2019 The Ohio State University
The Impacts of Physical Attacks Lifetime Vs. Attack arrival rate 2/24/2019 The Ohio State University
The Ohio State University Defense Strategies Over-provisioning sensor nodes Deploying more sensors to compensate the damage of blind attacks [9] Using sacrificial node to compensate the weakness of sensors in sensing capacity compared with the attacker [11] 2/24/2019 The Ohio State University
The Ohio State University Final Remarks The principles of Over Provisioning QOP: Quantitative Over Provisioning on network resources Practices of Over Provisioning in Overlay Networks Secure Overlay Forwarding Systems – Layers and Connectivity Resilient Structure P2P systems – Neighbor connectivity QoS aware Overlay multicast and anycast – Path Sensor networks Reliable sensor network – limited mobility sensor nodes Resilience to Physical attacks – node and structure 2/24/2019 The Ohio State University
The Ohio State University References S. Wang, Dong Xuan, R. Bettati and W. Zhao, “Providing Absolute Differentiated Services for Real-Time Applications in Static-Priority Scheduling Networks”, in IEEE/ACM Transactions on Networking (ToN), Vol 12, No. 2, April 2004. S. Wang, Dong Xuan, R. Bettati and W. Zhao, “Differentiated Services with Statistical Real-Time Guarantees in Static-Priority Scheduling Networks”, in Proc. of IEEE RTSS, 2001. S. Wang, Dong Xuan and W. Zhao, “On Resilience of Structured Peer-to-Peer Systems”, in Proc. of IEEE GLOBECOM, Dec. 2003. Dong Xuan, S. Chellappan and M. Krishnamoorthy, “RChord: An Enhanced Chord System Resilient to Routing Attacks”, in Proc. of IEEE ICCNMC, Oct. 2003. W. Jia, W. Zhao, Dong Xuan, and G. Xu, “An Efficient Fault-Tolerant Multicast Routing Protocol with Core-Based Tree Techniques”, in IEEE Transactions on Parallel and Distributed Systems (TPDS), Vol. 10, No. 10, Oct. 1999. Dong Xuan, W. Jia, W. Zhao, and H. Zhu, “A Routing Protocol for Anycast Messages”, in IEEE Transactions on Parallel and Distributed Systems (TPDS), Vol. 11, No. 6, June 2000. W. Jia, Dong Xuan, W. Tu, L. Lin and W. Zhao, “Distributed Admission Control for Anycast Flows”, in IEEE Transactions on Parallel and Distributed Systems (TPDS), Vol 15, No. 8, August 2004. Dong Xuan, S. Chellappan, X. Wang and S. Wang, ”Analyzing the Secure Overlay Services Architecture under Intelligent DDoS Attacks”, in Proc. of IEEE International Conference on Distributed Computing Systems (ICDCS), March 2004. Xun Wang, Wenjun Gu, Sriram Chellappan, Kurt Schosek, Dong Xuan, “Lifetime Optimization of Sensor Networks under Physical Attacks ”, IEEE ICC 2005. S. Chellappan, X. Bai, B. Ma and Dong Xuan, Mobility Limited Flip-based Sensor Network Deployment, accepted by IEEE Transactions on Parallel and Distributed Systems (TPDS), Oct. 2005. W. Gu, X. Wang, S. Chellappan, Dong Xuan and Ten H. Lai, Defending against Search-based Physical Attacks in Sensor Networks, to appear in Proc. of IEEE MASS, Nov. 2005 2/24/2019 The Ohio State University