On Proxy Server based Multipath Connections (PSMC) PhD Proposal Yu Cai 12/2003 University of Colorado at Colorado Springs

Outline 1. Introduction 1. Introduction 2. Related work 2. Related work 3. PSMC algorithms 3. PSMC algorithms 4. PSMC protocols 4. PSMC protocols 5. PSMC applications 5. PSMC applications 6. PSMC security 6. PSMC security 7. Conclusion 7. Conclusion

Introduction Single path connection vs. multipath connections Single path connection: most commonly-used network connection model in today’s network environment. Single path connection: most commonly-used network connection model in today’s network environment. Multipath connections: provide potentially multiple paths between network nodes. The traffic from a source can be spread over multiple paths and transmitted in parallel through the network. Multipath connections: provide potentially multiple paths between network nodes. The traffic from a source can be spread over multiple paths and transmitted in parallel through the network.

Why Multipath Connections Improve the network security by providing alternate paths Improve the network security by providing alternate paths Improve the network reliability, stability and availability Improve the network reliability, stability and availability Improve the network performance by increasing the aggregate bandwidth between network nodes Improve the network performance by increasing the aggregate bandwidth between network nodes Utilize the network resources more efficiently Utilize the network resources more efficiently Cope well with network congestion, link breakage, burst traffic and potential attacks Cope well with network congestion, link breakage, burst traffic and potential attacks Provide better quality-of-service Provide better quality-of-service

Related Works on Multipath Connections Multipath connections have been studied since 70s. Multipath connections have been studied since 70s. The IBM Systems Network Architecture (SNA) in 1974 The IBM Systems Network Architecture (SNA) in 1974 Nicholas F. Maxemchuk in 1975, the dispersity routing Nicholas F. Maxemchuk in 1975, the dispersity routing Classification of multipath connections based on OSI 7-layer model. Classification of multipath connections based on OSI 7-layer model. Physical layer: Multipath Interference; Antenna Array. Physical layer: Multipath Interference; Antenna Array. Data link layer: Link Aggregation, defined in IEEE 802.3ad. (requires additional hardware support) Data link layer: Link Aggregation, defined in IEEE 802.3ad. (requires additional hardware support)

Related Works on Multipath Connections Network layer: studied extensively as multipath routing. Network layer: studied extensively as multipath routing. Wired network. (requires changes on routers) Wired network. (requires changes on routers) Table-driven routing (link state or distance vector). MDVA(Multipath distance vector algorithm ) [VG01]; [Chen98] Table-driven routing (link state or distance vector). MDVA(Multipath distance vector algorithm ) [VG01]; [Chen98] Wireless ad hoc network. (only for ad hoc network) Wireless ad hoc network. (only for ad hoc network) On-demand routing. On-demand routing. SMR(Split Multipath Routing ) [LG00], Source Routing. Source Routing. MSR(Multipath Source Routing ) [ZZS+02] Transport layer: Linux multipath connections for multiple ISP connections. (no fail-over mechanism). Transport layer: Linux multipath connections for multiple ISP connections. (no fail-over mechanism).

Proxy Server based Multipath Connections (PSMC) Existing multipath connection approaches have various limitations and drawbacks. Existing multipath connection approaches have various limitations and drawbacks. We want new solution: We want new solution: Must be compatible with current network and don’t require changes on network infrastructure; Must be compatible with current network and don’t require changes on network infrastructure; Must be robust and reliable with high performance; Must be robust and reliable with high performance; Must be flexible when deployed so more applications can benefit from it. Must be flexible when deployed so more applications can benefit from it. We propose to study a new multipath connection approach: proxy servers based multipath connections (PSMC). We propose to study a new multipath connection approach: proxy servers based multipath connections (PSMC).

The Key Idea of PSMC The key ideas of PSMC is as followings. The key ideas of PSMC is as followings. By using a set of connection relay proxy servers, we could set up indirect routes via the proxy servers, and transport packets over the network through the indirect routes. By using a set of connection relay proxy servers, we could set up indirect routes via the proxy servers, and transport packets over the network through the indirect routes. By enhancing existing TCP/IP protocols, we could efficiently distribute and reassemble packets among multiple paths at two end nodes, and increase end-to-end TCP throughput. By enhancing existing TCP/IP protocols, we could efficiently distribute and reassemble packets among multiple paths at two end nodes, and increase end-to-end TCP throughput. The approach offers applications the ability to improve network security, reliability, performance, stability, availability and efficiency. The approach offers applications the ability to improve network security, reliability, performance, stability, availability and efficiency.

PSMC Diagram

Three Key Parts in PSMC The multipath sender: distributes packets over the selected multiple paths efficiently and adaptively. The multipath sender: distributes packets over the selected multiple paths efficiently and adaptively. The intermediate connection relay proxy servers: examine the incoming packets and forward them to the end server. The intermediate connection relay proxy servers: examine the incoming packets and forward them to the end server. The multipath receiver: collects the packets from multiple paths, reassembles them in order and delivers them to the user. The multipath receiver: collects the packets from multiple paths, reassembles them in order and delivers them to the user.

Why PSMC Compatibility: Utilizes existing TCP/IP protocols and network infrastructure. Don’t require changes on physical network infrastructure. Compatibility: Utilizes existing TCP/IP protocols and network infrastructure. Don’t require changes on physical network infrastructure. Flexibility: Can be more conveniently and adaptively deployed in various network environments. Flexibility: Can be more conveniently and adaptively deployed in various network environments. Usability: A large number of applications in various categories could benefit from utilizing PSMC. Usability: A large number of applications in various categories could benefit from utilizing PSMC. Reliability: Reliable and robust protocol with high end- to-end performance. Reliability: Reliable and robust protocol with high end- to-end performance.

Algorithms for PSMC Proxy server selection is a critical decision in PSMC. Different server selections result in different performance. Proxy server selection is a critical decision in PSMC. Different server selections result in different performance. Needs to solve the following two proxy server selection problems. Needs to solve the following two proxy server selection problems. 1) Server Selection Problem. Given the target server and a set of proxy servers, choose the best proxy server or servers for a client or for a group of clients, to achieve the maximum aggregate bandwidth. 2) Server Placement Problem. Given the target server and a set of network nodes, choose the best node(s) to place the proxy servers, to maximize the aggregate bandwidth.

Diagram of Sever Selection / Placement Problem How to avoid joint paths when selecting proxy servers? (joint path might become potential bottleneck) How to select geographically diverse proxy servers? Server selection problem Server placement problem

Related Work on Algorithms Mirror server and cache server selection problem has been studied recent years. Mirror server and cache server selection problem has been studied recent years. Formal approach: abstract network model; use graph theory. Formal approach: abstract network model; use graph theory. Common assumptions when getting network model: Common assumptions when getting network model: a) network topology is known, b) the cost associated with each path is known, c) single and static network connections. Algorithms include [QPV01]: Algorithms include [QPV01]: (selecting M replicas among N potential sites) tree-basedgreedyrandom hot spot O(N 3 M 2 )O(N 2 M)O(NM)N 2 + min (NlogN, NM)

Algorithms for Parallel Download Problem NP-hard problem. We plan to develop heuristic algorithms, or by loosing the optimal constrains to simplify the problem to make it solvable in P-time. NP-hard problem. We plan to develop heuristic algorithms, or by loosing the optimal constrains to simplify the problem to make it solvable in P-time. We have developed genetic algorithms to choose best mirror sites for parallel download from multiple mirror sites. The problem can be viewed as a sub problem of PSMC. We have developed genetic algorithms to choose best mirror sites for parallel download from multiple mirror sites. The problem can be viewed as a sub problem of PSMC.

Parallel Download Algorithm Performance Performance result of the parallel download algorithms tested on the simulated network and real-world network looks promising.

PSMC Protocols: Packets Handling Protocols need to be designed for packets handling: Protocols need to be designed for packets handling: Distribute / reassemble packets: add a thin layer between TCP and IP. Modify the Linux kernel. Distribute / reassemble packets: add a thin layer between TCP and IP. Modify the Linux kernel. Transmit packets: use IP Tunnel or IPSec to enable indirect routes. Transmit packets: use IP Tunnel or IPSec to enable indirect routes. Why adding a thin layer for packets distribution and reassembling? Why adding a thin layer for packets distribution and reassembling? Utilize existing TCP protocols, particularly the packets re- sequencing and re-sending mechanism. Utilize existing TCP protocols, particularly the packets re- sequencing and re-sending mechanism. Hide the complexity of multipath connections from end user. Hide the complexity of multipath connections from end user. Maintain the high end-to-end TCP throughput. Maintain the high end-to-end TCP throughput.

PSMC Protocols: IP Tunnel IP tunnel is a technique to encapsulate IP datagram within IP datagram. This allows datagram destined for one IP address to be wrapped and redirected to another IP address. IP tunnel is a technique to encapsulate IP datagram within IP datagram. This allows datagram destined for one IP address to be wrapped and redirected to another IP address. IPSec is an extension to the IP protocol which provides security to the IP and the upper-layer protocols. The IPSec architecture is described in the RFC2401. IPSec is an extension to the IP protocol which provides security to the IP and the upper-layer protocols. The IPSec architecture is described in the RFC2401. Why IP Tunnel: Why IP Tunnel: IP Tunneling is well developed and widely available. IP Tunneling is well developed and widely available. It is a layer 2 protocol, transparent to higher layer. It is a layer 2 protocol, transparent to higher layer. IP Tunneling performance is acceptable. IP Tunneling performance is acceptable. We have investigated other approaches including SOCKS proxy server and Zebedee, which don’t fit our needs. We have investigated other approaches including SOCKS proxy server and Zebedee, which don’t fit our needs.

Special Issues for PSMC Protocols Several special issues for PSMC protocols: Several special issues for PSMC protocols: Based on the feedback from end server, dynamically adjust packets distribution. Based on the feedback from end server, dynamically adjust packets distribution. Outgoing packets might contain redundant information and/or probing message. Outgoing packets might contain redundant information and/or probing message. Fail-over mechanism, packets resend and re-sequencing mechanism, when packets are lost or connections are broken. Fail-over mechanism, packets resend and re-sequencing mechanism, when packets are lost or connections are broken. Sticky-connection mechanism: when some packets need to be sent through a particular path. Sticky-connection mechanism: when some packets need to be sent through a particular path. Related work: Related work: ATCP (ad hoc TCP) [LS01]. ATCP (ad hoc TCP) [LS01]. Linux Virtual Server (LVS). Linux Virtual Server (LVS). Virtual Private Network (VPN) Virtual Private Network (VPN)

IP Tunnel and IPSec

PSMC Diagram

PSMC Applications Secure Collective Defense (SCOLD) network Secure Collective Defense (SCOLD) network PSMC in wireless ad hoc network. PSMC in wireless ad hoc network. Indirect route / additional bandwidth upon operational requests. Indirect route / additional bandwidth upon operational requests. QoS for video streaming. QoS for video streaming. Parallel download from multiple mirror sites. Parallel download from multiple mirror sites.

Secure Collective Defense (SCOLD) network SCOLD tolerates the DDoS attacks through indirect routes via proxy servers, and improves network performance by spreading packets through multiple indirect routes. SCOLD tolerates the DDoS attacks through indirect routes via proxy servers, and improves network performance by spreading packets through multiple indirect routes. SCOLD will incorporate various cyber security techniques, like secure DNS update, Autonomous Anti-DDoS network, IDIP(Intrusion Detection and Isolation Protocol) protocols. SCOLD will incorporate various cyber security techniques, like secure DNS update, Autonomous Anti-DDoS network, IDIP(Intrusion Detection and Isolation Protocol) protocols. The prototype of SCOLD system version 1.0 is finished with secure DNS update and indirect route. The prototype of SCOLD system version 1.0 is finished with secure DNS update and indirect route. We plan to enhance SCOLD for better scalability, reliability, performance and security. We plan to enhance SCOLD for better scalability, reliability, performance and security.

SCOLD: victim under DDoS attacks Victim aaaabbbbcccc A.comB.comC.com... A B C R R2 R1R3 Back door: Alternate Gateways DNS DDoS Attack Traffic Client Traffic Main gateway R under attacks, we want to inform Clients to go through the “back door” - alternate gateways R1- R3. We needs to hide IPs of R1-R3, otherwise they are subject to potential attacks too. how to inform Clients? how to hide IPs of R1-R3? target.com DNS1 DNS2DNS3

SCOLD: raise alarm (1) and inform clients (2) 1. IDS on gateway R detects intrusion, raise alarm to Reroute Coordinator. 2. Coordinator informs clients for new route: a) inform clients’ DNS; b) inform clients’ network proxy server; c) inform clients directly; d) inform the proxy servers and ask the proxy server do (a – c). Victim aaaabbbbcccc A.comB.comC.com... A B C R R2 R1 R3 DNS target.com DNS1 DNS2DNS3 Reroute Coordinator 1: raise alarm 2: inform clients Proxy1

SCOLD: set up new indirect route (3) Victim aaaabbbbcccc A.comB.comC.com... A B C R R2 R1R3 DNS target.com DNS1 DNS2DNS3 Reroute Coordinator 3: new route Proxy1Proxy2 3. Clients set up new indirect route to target via proxy servers. Proxy servers: equipped with IDS to defend attacks; hide alternate gateway and reroute coordinator; provide potential multiple paths. Proxy3

SCOLD Testbed

No DDoS attack direct route DDoS attack direct route No DDoS attack indirect route DDoS attack indirect route 0.49 ms225 ms0.65 ms Preliminary result of SCOLD Table 1: Ping Response Time (on 3 hop route) Table 2: SCOLD FTP/HTTP download Test (from client to target) Table 2: SCOLD FTP/HTTP download Test (from client to target) Table 3: Time to Set up Indirect Route in SCOLD Table 3: Time to Set up Indirect Route in SCOLD PingLess than 1 s HTTPLess than 1 s FTPLess than 1 s

PSMC Applications Evaluation The performance and overhead of multipath connections will be evaluated. The performance and overhead of multipath connections will be evaluated. PSMC will be compared with other multipath connection approaches, like source routing, and Linux multipath connections. PSMC will be compared with other multipath connection approaches, like source routing, and Linux multipath connections. Extensive simulation study on PSMC applications in virtual network, real network, small scale network and large scale network will be conducted. Extensive simulation study on PSMC applications in virtual network, real network, small scale network and large scale network will be conducted.

Security Issues Related to PSMC Potential security issues raised by misusing of PSMC: how to control aggressive clients? Potential security issues raised by misusing of PSMC: how to control aggressive clients? Potential attacks against PSMC: Tunneling to death? (similar to ping to death). Potential attacks against PSMC: Tunneling to death? (similar to ping to death). How to detect and deal with comprised nodes in PSMC network? How to detect and deal with comprised nodes in PSMC network? Study the collective defend mechanism to tie different organizations with better cooperation and collaboration. Study the collective defend mechanism to tie different organizations with better cooperation and collaboration.

Research Plan Will systematically study PSMC in the following areas: Will systematically study PSMC in the following areas: Algorithms for server selections Algorithms for server selections Protocols for packet handling Protocols for packet handling Applications Applications Security issues Security issues

Thank you!