A network that expands beyond a metropolitan area is a widearea network. Wideareanetworks share a few characteristics with local areanetworks: they interconnect computers, they use some form of media for the interconnection, and they support network applications. More importantly, however, wideareanetworks differ from local areanetworks in a number of ways. Wideareanetworks include both data networks, such as the Internet, and voice networks, such as telephone systems, whereas local areanetworks in almost all cases include only data networks. Wideareanetworks can interconnect thousands, tens of thousands, or more workstations so that any one workstation can transfer data to any other workstation. As the name implies, wideareanetworks can cover large geographic distances, including the entire earth.
moot the choice of packet queueing discipline.
Finally, we configure the routers to have infinite buffer space. Packet and cell queues are thus allowed to grow without bound and consequently routers never drop data. Similarly, by default, VCSIM switches never drop cells. Wide-areanetworks avoid dropping data because the losses trigger retransmissions in the transport-level protocols running in the hosts. Retransmis- sions in turn lead these protocols to slow the rate at which they send packets, which results in lower throughput for the hosts. As this process acts on all traffic sources experiencing data loss, network utilization may decrease even as offered load increases. Networks that avoid these resource inefficiencies are considered stable. Naturally, we cannot build real networks with infinite buffers. However, simulating this scenario lets us measure how much memory would be needed in order not to drop data. Knowledge of this upper bound on memory sizes is very useful for wide-area network design.
Online data storage has become within reach of many users and organizations over the last number of years. The exponential increase of Internet bandwidth is an important enabler for this trend. SURFnet is investigating various technologies that can be used to create an open storage architecture for higher education and research in the Netherlands, by leveraging its high-speed state-of-the-art national network infrastructure. This effort focuses primarily on providing high-performance and highly available online storage facilities to its participants in the SURFnet WideArea Network (WAN). As part of this effort, SURFnet has executed a scan on existing solutions and technologies for storage in WideAreaNetworks (WANs) (see the survey document at ). The Distributed Replicated Block Device (DRBD)  is identified as an interesting storage technology that may be used for WAN data replication. In this document, we focus on DRBD usage in WANs: in parallel, we are evaluating other technologies identified as interesting by the scan (GPFS and Gluster).
For more information about available SOLUTIONware configurations, visit the Brocade SAN Solution Center at www.brocade.com/san.
Connecting SANs Over WideAreaNetworks (WANs)
To connect areas that surpass the reach of ELWL GBICs, link extenders, and DWDM, protocols.The following case studies demonstrate how organizations can deploy Remote Device Mapping and Remote Switch configurations over existing WAN infrastructures.
To accomplish the safety-critical mission of transmission and distribution automation, high avail- ability and stability are always required in the industrial communication networks. The IEC 62439 Standard (Industrial communication networks-High availability automation networks) Part 3 de- fines the Parallel Redundancy Protocol (PRP) and High-availability Seamless Redundancy (HSR) protocols. Both redundancy protocols target the process bus described in IEC61850 9-2  and support only Local AreaNetworks (LANs). It has become an active research area to extend the re- dundancy protocols to the WideArea Network (WAN) communication with existing network in- frastructure. This paper offers a novel solution, namely the parallel redundancy protocol over the wideareanetworks (PRPW), to enable PRP over WANs with no added overheads to the data frame, while retaining full compatibility. Moreover, PRPW also strengthens the cyber security against spoofing attacks by maintaining a comprehensive list of remote communication units.
Shuhao Liu and Baochun Li
Abstract: Software-Defined Networking (SDN) has emerged as a promising direction for next-generation network design. Due to its clean-slate and highly flexible design, it is believed to be the foundational principle for designing network architectures and improving their flexibility, resilience, reliability, and security. As the technology matures, research in both industry and academia has designed a considerable number of tools to scale software-defined networks, in preparation for the wide deployment in wide-areanetworks. In this paper, we survey the mechanisms that can be used to address the scalability issues in software-defined wide-areanetworks. Starting from a successful distributed system, the Domain Name System, we discuss the essential elements to make a large scale network infrastructure scalable. Then, the existing technologies proposed in the literature are reviewed in three categories: scaling out/up the data plane and scaling the control plane. We conclude with possible research directions towards scaling software-defined wide-areanetworks.
In this paper ,we make four main contribu tions. We confirm that architecture and the producer-consumer problem can connect to sur- mount this obstacle. Continuing with this rationale,we probe how wide-areanetworks can be applied to he improvement of semaphores. Next,we understand how spreadsheets can be applied to the evaluation of super pages.Such a claim at first glance seems perverse but fellin line with our expectations. In the end,we use homogeneous
TCP congestion control can perform badly in highspeed wideareanetworks because of its slow response with large congestion win- dows. The challenge for any alternative protocol is to better utilize networks with high bandwidth-delay products in a simple and ro- bust manner without interacting badly with existing traffic. Scal- able TCP is a simple sender-side alteration to the TCP conges- tion window update algorithm. It offers a robust mechanism to improve performance in highspeed wideareanetworks using tradi- tional TCP receivers. Scalable TCP is designed to be incrementally deployable and behaves identically to traditional TCP stacks when small windows are sufficient. The performance of the scheme is evaluated through experimental results gathered using a Scalable TCP implementation for the Linux operating system and a gigabit transatlantic network. The results gathered suggest that the deploy- ment of Scalable TCP would have negligible impact on existing network traffic at the same time as improving bulk transfer perfor- mance in highspeed wideareanetworks.
In recent years it has been shown that database replication is promising in improving performance and fault toler- ance of database systems. Data replication means that there exist many copies of the same data. A challenge is replica control, i.e., to keep copies consistent despite updates. Many replica control protocols have been proposed. Most of these protocols have two shortcomings. Firstly, although these protocols perform well for cluster based systems in Local AreaNetworks (LAN) they are not applicable in WideAreaNetworks (WAN) due to the much longer commu- nication delay in WAN. Secondly, database replication must gurantee certain levels of transaction isolation, i.e., to what extent transactions will interfere with each other. Most of the existing protocols guarantee serializability, which has become less popular than Snapshot Isolation (SI), a new transaction isolation level. My thesis aims to propose a replication solution which guarantees SI and works well for WANs. Furthermore, I use a middleware approach which provides replica control outside the database system. This provides flexibility and allows heterogeneous config- urations. Existing middleware based approaches have severe restrictions. For instance, some require all operations of a transaction to be known in advance or that transactions must be marked as read-only or update. Furthermore, concurrency control is usually at a coarse level, e.g., table level. My research aims in overcoming these restrictions in order to provide a flexible and transparent solution.
In this paper, we introduce a replication-based approach to achieve fast and reliable stream processing over wideareanetworks. Our technique provides replica transparency, guaranteeing that replicated processing, regardless of fail- ures and congestions, will produce what non-replicated pro- cessing would produce if all the system components are free from failure and latency. In our approach, replicas compete with each other to make the earliest impact, while guaran- teeing that they will eventually produce the same collec- tion of tuples. These replicas, however, can run in differ- ent orders so as to avoid the overhead of other previous ap- proaches. Our technique can merge replicated streams into a single stream, while filtering out duplicates. If required by the end application, it can also be instructed to restore the order of the stream that non-replicated processing would produce.
Ethernet service networks bypass the PSTN; companies offering Ethernet/IP packet networks have laid their own gigabit Ethernet fiber optic networks in large cities. When an organization signs up for service, the packet network company installs new fiber optic cables from their city- wide WAN backbone into the organization's office complex and connect it to an Ethernet switch. The organization simply plus their network into their Ethernet switch and begins using the service. All traffic entering the packet network must be Ethernet using IP. Since most organizations today use Ethernet and IP in the LAN and BN environment, Ethernet/IP avoids the need to translate or encapsulate to generate addresses for LAN or BN traffic and gains in throughput. It avoids complexity, meaning that companies do not have to add staff knowledgeable in the different WAN protocols, software, and hardware these technologies require. This technology is offered by relatively new startup companies like Yipes.com.
Although MANs are WANs, their relatively short distance span means that the cost per bit transmitted is lower than it is in national and international WANs.
Consequently, typical transmission speeds are faster. If you have a smartphone or tablet with 3G or 4G cellular access, then you already use a MAN. Cellular networks almost always span a single MAN or even a single city. However, we will see that wired MANs are important for corporations because site-to-site traffic is large and is more efficiently transmitted over wires.
have equal standing within the network. Nodes are “devices on a network that demand or supply services or where transmission paths are connected” .
1.5 Structure of the Thesis
This document firstly presents a literature review in the area of service discov- ery. Using a well defined set of requirements for service discovery techniques, a set of diverse problems is described. Designed to tackle this set of problems, a centralised approach to service discovery is presented in Chapter 3 using the Session Initiation Protocol (SIP). This solution, named the SIP Service, is shown to facilitate importing devices and services from one domain to another. Addi- tionally, the SIP Service demonstrates protocol bridging between SIP and the UPnP protocol. After a considered analysis of this approach, the SIP Service is found to scale inefficiently. Consequently, in Chapter 4, an alternative and more suitable solution to the service discovery problem is identified using P2P service overlays.
Roaming: Wandering out of range or between net-
work segments confuses applications and protocol stacks. Most wireless users today feel that they are using "bleeding edge" technology, so they accept frequent disruptions in service because they seem to be a fact of life. NetMotion's advanced roaming algorithms allow users to roam—not just from one network interconnect to another, but across net- work segments, and even to different networks altogether without disconnecting. NetMotion users don't have to restart their machines to obtain a new address. Users maintain application persistence even when they cross an IP subnet boundary. Employing its client-server architecture, the Net- Motion server maintains the state of each mobile device and handles the complex session manage- ment required for continuous connections to the network-based applications. When a mobile device becomes unreachable because it suspends opera- tion, moves out of coverage, or changes its current "point of presence" address (roaming), the server maintains the connection to the network peer by acknowledging receipt of data and queuing requests. Once the mobile node and the NetMotion server are in contact again, the state of each exist- ing connection is resynchronized, thus providing continuous network computing—all without mod- ification to applications and network infrastruc- tures.
Railway communication systems for applications within the signals and control systems, telecommunications, electrical, and rolling stock disciplines are increasingly based on ethernet and internet protocol (IP) enabled computer systems. Local, metropolitan, and widearea wired networks shall align with national and international standards to create open industry involvement, increased competition, and optimal asset stewardship outcomes.
Chapter 7. Introduction to the OPS-based widearea network 99
; it selects the FDL with the minimum gap only between the last queued packet and the new one.
Despite the fact that the WDS algorithm can be relatively simple to implement, taking per-packet decisions requires too much computations considering that each switch has several ports, each port several wavelengths and each wavelength trans- ports packets at 10 Gbit/s or above. To overtake this problem, OPS concepts are recently extended to a connection-oriented network scenario , for instance based on MPLS. In this scenario, a suitable design of WDS algorithms permits to obtain fairly good performance, by exploiting queuing behaviors related to the connection- oriented nature of the traﬃc, but with a signiﬁcant saving in term of processing eﬀort for the switch control with respect to the connectionless case.
Fig. 5. Illustration of a WuR-enabled IoT testbed with indirect 3GPP connection where the mobile phone serves as a relay for cellular connection.
C. Insights from IoT Projects
In addition to the aforementioned examples, there have been many undergoing IoT projects targeted at different applications, such as smart agriculture, disaster detection, and digital health, which are funded by both industrial and academic organizations. The aforementioned implementation example of LoRa is a five-year project on smart city with industry and academic collaborations 2 . Various projects with similar purposes have been supported by different companies, such as Intel and IBM, and they have been carried out in different cities around the world to show the potential of deploying LPWA networks in a large scale. Another well-known example of IoT project on smart agriculture is that Fujitsu has launched a company in Finland to produce and sell vegetables year-round with artificial-light plant factory 3 . In order to make up the shortage of leaf vegetables in northern, sensors have been deployed in the greenhouse to collect information and a cloud server has been built up to realize automatic control of environment in the greenhouse. All these IoT projects give us the confidence to deploy LPWA networks in the real world.