Machine-Learned Recognition of Network Traffic for Optimization through Protocol Selection

We introduce optimization through protocol selection (OPS) as a technique to improve bulk-data transfer on shared wide-area networks (WANs). Instead of just fine-tuning the parameters of a network protocol, our empirical results show that the selection of the protocol itself can result in up to four times higher throughput in some key cases. However, OPS for the foreground traffic (e.g., TCP CUBIC, TCP BBR, UDT) depends on knowledge about the network protocols used by the background traffic (i.e., other users). Therefore, we build and empirically evaluate several machine-learned (ML) classifiers, trained on local round-trip time (RTT) time-series data gathered using active probing, to recognize the mix of network protocols in the background with an accuracy of up to 0.96.

Performance evaluation of mixed SCTP and TCP traffic over last hop wifi

The use of the internet has increased significantly with the continued increase in wireless communication devices. Recently, there is a large number of research contribution focused on Stream Control Transmission Protocol (SCTP). Multi-homing is an important feature of SCTP which improves the communication performance by usage of multiple paths during association establishment, and it can bring significant improvements of throughput. In this thesis we evaluate the performance of SCTP and TCP traffic in the WLANs and we investigate the SCTP multi-homing to improve the communication performance in WLANs. We conducted some experiments to evaluate the performance of SCTP multi-homed host under various channel bit rates and mobility speeds. The results indicate that when the intensity of background traffic increases the SCTP multi-homed host with higher channel bit rate has better performance. In addition, the SCTP multi-homed host with using lower mobility speed has higher performance (throughput, delay and packet loss).

Performance evaluation of mixed SCTP and TCP traffic over last hop wifi

The use of the internet has increased significantly with the continued increase in wireless communication devices. Recently, there is a large number of research contribution focused on Stream Control Transmission Protocol (SCTP). Multi-homing is an important feature of SCTP which improves the communication performance by usage of multiple paths during association establishment, and it can bring significant improvements of throughput. In this thesis we evaluate the performance of SCTP and TCP traffic in the WLANs and we investigate the SCTP multi-homing to improve the communication performance in WLANs. We conducted some experiments to evaluate the performance of SCTP multi-homed host under various channel bit rates and mobility speeds. The results indicate that when the intensity of background traffic increases the SCTP multi-homed host with higher channel bit rate has better performance. In addition, the SCTP multi-homed host with using lower mobility speed has higher performance (throughput, delay and packet loss).

Analisis Performansi Routing OSPF menggunakan RYU Controller dan POX Controller pada Software Defined Networking

Software Defined Networking (SDN) adalah teknologi baru yang dikembangkan untuk mengatasi masalah kompleksitas konfigurasi jaringan dengan pengelolaan lebih terpusat karena memisahkan antara control plane dan data plane. Penelitian ini menganalisis perbandingan performansi SDN menggunakan ryu controller dan pox controller. Protocol yang digunakan adalah protocol TCP dan UDP sedangkan routingnya menggunakan OSPF. Penerapan algoritma dijkstra dari perutean OSPF ke topologi fat tree pada SDN akan diukur unjuk kerjanya berdasarkan parameter Quality of Service yaitu delay, jitter, dan packet loss pada skenario tanpa background traffic. Transmisi data menggunakan traffic protocol TCP dengan RYU controller dan protokol UDP lebih baik karena memiliki delay sebesar 49.44% dan delay yang lebih stabil sebesar 0.01 %. Jitter yang dihasilkan adalah 27.59% lebih baik daripada POX controller sebesar 72.41% dan untuk protokol UDP menggunakan POX controller 99.97% sedangkan traffic protocol UDP menggunakan RYU controller lebih baik sebesar 0.03%. Packet loss dari kedua controller didapatkan hasil protokol TCP sangat bagus sebesar 0% sedangkan protokol UDP menggunakan POX controller 83.13%, menggunakan RYU controller 16.87%.

ANALISA PERFORMANSI CONTROLLER PADA ARSITEKTUR JARINGAN SOFTWARE DEFINED NETWORK (SDN)

AbstrakSoftware Defined Network (SDN) merupakan sebuah konsep pendekatan baru dalam jaringan untuk mendesain, membangun serta mengelola suatu jaringan komputer. Konsep ini melakukan pemisahan terhadap Data Plane dan Control Plane. Dalam konsep SDN ini terdapat suatu komponen penting yang bertanggung jawab terhadap segala aturan dalam pengelolaan dan pendistribusian informasi terhadap seluruh perangkat jaringan yaitu Controller. Karena peran Controller yang penting maka performa dari Controller perlu diuji sehingga dapat mengetahui kemampuan dari Controller yang digunakan. Dalam penelitian ini dilakukan perbandingan analisis nilai Quality of Services (QoS) terhadap implementasi SDN menggunakan Controller Floodlight dan Ryu dengan menjalankan topologi linear dan mesh dalam jumlah Switch yang beragam mulai dari 4, 8, 12 dan 16 Switch. Selama pengujian berlangsung dari node sumber ke node tujuan yang sama juga dialiri variasi background traffic mulai dari 50 hingga 200 Mbps. Hasil yang didapatkan yaitu Controller Ryu memiliki nilai QoS yang lebih baik dari floodlight pada semua topologi yang diujikan, nilai latency dan jitter pada floodlight lebih tinggi dari ryu serta cenderung meningkat pada traffic 100 Mbps Pada throughput, ryu memiliki nilai lebih tinggi dengan kisaran 856-933 Kbps. Sedangkan pada packet loss floodlight lebih tinggi sementara ryu hanya memiliki rata-rata packet loss sebesar 0,5%. Namun pada pengujian hanya pada jumlah switch, floodlight menjamin dalam tingkat respons serta pengelolaan data yang besar di dalam arsitektur jaringan SDN.Abstract Software Defined Network (SDN) is a concept of a new approach in networking to design, build and manage a computer network. This concept separates the Data Plane and Control Plane. In this SDN concept there is an important component that is responsible for all rules in the management and distribution of information to all network devices that is Controller. Due to the important Controller role then the performance of the Controller needs to be tested so as to know the ability of the Controller to use. In this study, a comparison of Quality of Service (QoS) value analysis on SDN implementation using Floodlight and Ryu Controller by running linear and mesh topology in varying number of Switches ranging from 4, 8, 12 and 16 Switch. During the test from the source node to the same destination node is also varies background traffic ranging from 50 to 200 Mbps. The result is that Controller Ryu has better QoS value than floodlight on all tested topologies, the latency and jitter values on the floodlight are higher than ryu and tend to increase on 100 Mbps traffic. Throughput On ryu. have a higher value with the range of 856-933 Kbps. While the packet loss floodlight higher while ryu only have an average packet loss of 0.5%. But on testing only on the number of switches, the floodlight guarantees great response rates and data management within the SDN network architecture.