ATM'de lan hizmetleri
ATM-lan services
- Tez No: 66559
- Danışmanlar: PROF. DR. GÜNSEL DURUSOY
- Tez Türü: Yüksek Lisans
- Konular: Elektrik ve Elektronik Mühendisliği, Electrical and Electronics Engineering
- Anahtar Kelimeler: Belirtilmemiş.
- Yıl: 1997
- Dil: Türkçe
- Üniversite: İstanbul Teknik Üniversitesi
- Enstitü: Fen Bilimleri Enstitüsü
- Ana Bilim Dalı: Elektronik ve Haberleşme Mühendisliği Ana Bilim Dalı
- Bilim Dalı: Belirtilmemiş.
- Sayfa Sayısı: 110
Özet
ÖZET ATM şebekesi, kullanıcılara, veri, ses ve görüntüyü taşıma imkanı veren yüksek band genişlikli bir şebekedir. ATM, hem LANlarda hem de WANlarda kullanılabilen ve sabit uzunluklu hücreleri sanal devreler üzerinden ileten bağlantı yönelimli bir teknolojidir. Bu çalışmada, ATM teknolojisi hakkında genel bilgilerin verilmesinin yarımda ATM'in varolan LAN tabanlı protokoller ile birlikte çalışmasına yönelik LAN emülasyonu, doğal tarz protokoller ve MPOA üzerine yapılan çalışmalar incelenmiştir. LAN emülasyonunun amacı, varolan LAN istemcilerine, herhangi bir değişiklik gerektirmeden ATM omurga üzerinden veri yollama ve ATM tabanlı kaynaklar ile haberleşme izni vermektir. LAN emülasyon hizmeti, emüle edilmiş LAN davranışı göstererek, LAN istemcilerini ATM şebekeden gizler. Böylece, LAN tabanlı cihazlar, LAN tabanlı protokolleri kullanmaya devam eder. LANE, ATM' i şebeke katmam protokollerinden saklar bu nedenle ATM üzerinde koşan şebeke katmam protokoller ATM'nin QoS özelliğinden faydalanamazlar. Bu eksikliğin üstesinden gelebilmek için İP protokolü üzerinde bir çalışma yapılmıştır. Eğer, ATM, İP için sadece bir iletim ortamı olarak düşünülürse, daha etkili hizmet sağlayabilmek üzere IP, ATM üzerinde koşabilecek şekilde uyarlanabilir. IETF (Internet Engineering Task Force), ilk olarak bir çok şebeke veya veri bağlantı katmam protokollerini bir tek ATM bağlantısı (AAL5) üzerinden iletmeyi ve aym bağlantı üzerinden çoğullama yapmayı tanımlayan bir yöntem üzerinde çalışmıştır. Bu protokol, ATM üzerinde klasik İP olarak bilinir ve mantıksal İP alt şebekeler düşüncesi üzerine kuruludur. Klasik İP protokolü, hedefin kaynak düğümün bulunduğu İP alt şebekenin dışında olması durumunda, paketin yöneltilmesi için gerekli olan,“varsayılan yönlendirici”ihtiyacım ortadan kaldıramamaktadır. IETF içindeki ROLC çalışma grubu bu sınırlamanın üstesinden gelebilmek için NHRP protokolü üzerinde çalışmaktadır. NHRP, klasik İP modeli üzerine kurulmuştur ancak bu protokolde LIS kavramının yerine mantıksal NBMA şebekesi kavramı yer almaktadır. NBMA şebekesi, ATM, FR veya X.25 benzeri bir şebeke teknolojisidir ve bir çok cihaza aynı şebekeye bağlanma izni verir. Ancak bu protokol otomatik konfigürasyon ve yayın/çok-hedefli yayın desteğine sahip değildir. Benzer olarak RFC 1577 içinde de çok hedefli yayın için yeterli bir destek yoktur. Bunun için RFC 1112 içinde bu desteğe yönelik tanımlamalar yapılmıştır. Bu çalışmanın NHRP ve MPOA gibi protokoller içinde çok hedefli yayın için kullanılması beklenmektedir. MPOA, ikinci katmanda çalışan LANE protokolünün geliştirilmiş hali olarak düşünülebilir. MPOA, ilk üç katmanda çalışmakta ve ATM bulutu içinde herhangibir hedefe direkt bağlantı sağlamakta ve ATM'in QoS özelliğini de LANE'ye nazaran daha etkili bir biçimde kullanmaktadır. vın
Özet (Çeviri)
SUMMARY ATM-LAN SERVICES ATM offers users a high bandwith network, able to carry mix of data, voice and video traffic, and provides a unified LAN/WAN networking model. ATM can deliver on these promises because it is based on the connection-oriented transfer of data in small, fixed-sized cells, which are indivually switched across virtual circuits. Demand for high speed networks is being driven by a number of factors, all of which fuel the need for greater bandwith. These include: I. Distributed processing II. Client server applications III.Graphics based applications such as animation IV.Multimedia applications which combine data with voice and video ATM can offer extremely high bandwiths with very low letancy because of cell switching technology. ATM provides a service which combines the best features of circuit switching and packet switching. ATM have the capability to handle data, voice and video. ATM promises a single architecture for both LANs and WANs. ATM provides a scaleable architecture, which means that the same technology can be used to provide a 25 Mbps service to the desktop or a 622 Mbps carrier service. Today, the underlaying optical fiber technology is being tested at speeds up to 4.8 Gbps. The switched technology allows for the Quality of Service provided by the connection to be negotiated when the service is established. Furthermore, network resources are allocated on a connection basic, and thus can be better allocated and managed. Since ATM operates at the lower two layers of the OSI model (the Physical and Data Link layers) it is protocol independent. ATM provides simplified configuration and management as compared to router based networks. A simplified network architecture using the concept of Virtual LAN's will be the model of the future. Routers rather than being replaced, will be reassigned to provide routing between the Virtual LANs and sub-ATM speed WAN internetworking. Improved security is provided by the fact that switched, rather than shared, connections are utilized and through the added functionality of Virtual LANs and Virtual Routing. An ATM conversation has three stages: call set-up, data transfer, and call termination. The first step in any ATM conversation is to establish a connection between the participants which carries the traffic and supports the necessary characteristics. This is done by using signalling to request the ATM network to set up a connection and negotiate the connection's attributes. IXATM connections are point-to-point connections called Virtual Circuits. There are two types of connections that can be set up between sender and receiver: Permanent Virtual Circuits and Switched Virtual Circuits. A Permanent Virtual Circuit, or PVC, is a circuit which exists permanently in a network. PVCs are analogous to leased data lines. Switched Virtual Circuit, or SVC, is a circuit which exists only for the duration of a call. This type of service is analogous to that provided during an ordinary telephone call. The concept of establishing a connection which supports specific traffic charracteristics is explaned by ATM Classes of Service and Quality of Service. Four fundemental classes of traffic are supported by ATM I. Class A provides Circuit Emulation type service, such as that required by video or voice types of traffic II. Class B provides a variable bit rate voice or video service, such as that required by video conferencing or compressed voice or video applications IILClass C provides a connection-oriented data service IV.Class D provides a connectionless data transfer service To completely specify a desired service, we need to not only specify the service types, but also the Quality of Service (QoS) required. These QoS transmission parameters can be grouped into three main attributes: Throughput, Delay and Accuracy. ATM uses two types of network interfaces: a User to Network Interface (UNI) and a Network to Network Interface (NNI). The UNI defines the interface between the user and the network. There are two types of UNI interfaces: a private UNI,used when an ATM user connects to an ATM switch on the same corporate network, and a public UNI, used when an ATM user or network connects to a public service provider's ATM network. The NNI, or Network to Network Interface, defines the interface between two ATM switches in a public ATM network. Once the ATM connection is established, data can be sent over it by segmenting it into ATM cells and transferring the cells along the Virtual Circuit making up the connection. The ATM model actually sees Virtual Circuits as being composed of Virtual Paths and Virtual Channels. Virtual Paths and Virtual Channels are the mechanizm used to transport and route cells in ATM. A Virtual Channel (VC) is a logical transmission path which is used to transport cells between two end points. Each end point uses a Virtual Channel Identifier (VCI) to identify the transmission path. A Virtual Path (VP) is a group of Virtual Channels that share a common transmission path and have the same Virtual Path Identifier (VPI). An ATM sender-to-receiver connection or Virtual Circuit, is a particular Virtual Channel inside a particular Virtual Path. All ATM data transfer is done in the form of fixed-sized cells, consisting of 53 bytes. Of these 53 bytes, 48 bytes are the Cell Payload, into which the user data to be transferred is placed, and 5 bytes are the Cell Header.The structure of the cell payload depends on the ATM Adaptation Layer (AAL) used, and the Cell Header carries the information required to allow the network to switch it along the appropriate Virtual Circuit. After having an ATM connection to transfer data across, the user's data must be put in the payload of a sequence of ATM cells. This task is carried out by the ATM Adaptation Layer. The ATM Adaptation Layer is composed of two sub-layers. The Convergence Sublayer (CS) determines the way user data is to be segmented into cells (according to the type of service required), while the Segmentation And Reassembly (SAR) sublayer carries out the actual segmentation of user data into ATM cells and reconstruction of user data from incoming ATM cells. The cells is tranmitted using a particular physical method. The crucial point is that the ATM network needs to quickly and efficiently switch the cells at each switch along the Virtual Circuit that the cells at each switch along the Virtual Circuit that the cells are to follow. This is done by the ATM layer. ATM switching is done cell by cell, according to the information in the cell's header. An ATM switch may use just the VPI part of the cell's header to decide how to forward the cell (Virtual Path Switching), or just the VCI part (Virtual Channel Switching), or both. The ATM switch receives a cell on a particular mcoming port, and the cell is marked as belonging to a particular Virtual Circuit. The switch than examines its routing table, from which it finds out: I. On which outgoing port to forward the cell II. To what values the VPI/VCI need to be set on the outgoing cell After the data transfer is concluded, the ATM client again uses signalling addressed to the network, to tell the network the conversation is complete, and that the virtual circuit providing the connection can be released. LANE The purpose of LAN emulation is to allow existing LAN clients to send data over ATM backbones, and to communicate with ATM-based resources, without requiring any change in the LAN-based clients. The LAN emulation service essentially hides the ATM network from the LAN clients by emulating LAN behaviour, so that LAN-based devices can continue to use LAN based protocols. LANs assume that a packet sent over the LAN will be seen by all stations belonging to that LAN. This allows a station to generate a broadcast by sending a single packet, with an all destination address. In ATM networks, all connections are point-to-point, so that a packet generated by a station will only get to a single recevier. To reconcile these different models, the LAN emulation service has to generate copies of the packet to be broadcast, and send these copies to each member of the LAN being emulated. Since we want LAN-based stations to keep working without changes, it follows that they must be able to use LAN type addresses to identify destinations. Since ATM networks use ATM addresses to identify stations, and VPI/VCI tags to identify connections, the ATM LAN Emulation service will have to provide for address translation and resolution between the different forms. XILAN Emulation is a Layer 2 service and is independent of upper layer protocols, supporting both routable and non-routable protocols. The LAN Emulation User to Network Interface or LUNI as defined by the ATM Forum, specifies how Ethernet or Token Ring attached workstations connect to their counterparts over an ATM network. It also defines how ATM attached servers communicate with devices on existing Ethernet or Token Ring LANs. FDDI must be converted to either Ethernet or Token Ring to be transported over ATM. Emulated LANs do not mix media; a router is required to provide this level of connectivity. For example, an Ethernet client cannot communicate to a Token Ring server, this connectivity must be provided by a router. LUNI supports multiple Emulated LANs on the same physical LAN, but routers are required to provide connectivity between Emulated LANs. LAN Emulation allows existing LAN protocols, such as IP, IPX, AppleTalk, etc., to run over ATM networks without requiring any changes to the applications. LAN Emulation has a number of components, the LAN Emulation Client or LEC, the LAN Emulation Configuration Server or LECS, the LAN Emulation Server or LES, and the Broadcast and Unknown Server or BUS. An emulated LAN provides the functionary of a single Ethernet segment or Token Ring. LAN Emulation supports Unicast, Multicast and Unknown traffic types through the mechanisms provided by the LEC, LECS, LES and BUS. The main function of the LAN Emulation Client is that of address resolution, that is mapping MAC addresses to ATM addresses. The LEC interfaces to the ATM network over a LAN Emulation UNI. The LAN Emulation Configuration Server provides a LEC with configuration information, and the address of the LES. There is one LECS for all emulated LAN's, but each emulated LAN has its own LES and BUS. The LAN Emulation Server is responsible for registering and resolving MAC addresses to ATM addresses. The Broadcast and Unknown Server handles the function of broadcasting and multicasting over the ATM network. NATIVE MODE PROTOCOLS There are many networks using IP as the network layer protocol, and they use a variety of transport methods. For example Ethernet, Token Ring, FDDI, Dial-Up lines, etc. In each case, the IP implementation is aware of the transport mechanism used, and uses this knowledge to provide an efficient service. If we view ATM as just another Transport mechanism for IP to run over, we can customize the IP standart for transport over ATM to provide an efficient implementation. In particular, if IP is customized for ATM transport, then it can use ATM addresses for the link-level addresses. This is best understood by looking at the operation of the IP Address Resolution Protocol (ARP) on a traditional Ethernet LAN and on an ATM network. In an Ethernet LAN, the sender generates an ARP broadcast essentially asking IP address of receiver. This broadcast is received by all LAN stations, and the Xllintended destination responds with an ARP response containing its link-level (MAC) address. The sender then uses this address to send the data. In the case of IP over ATM, the sender sends a packet to an ARP server (not broadcast) asking for the destination's link-level address. The server returns the destination's ATM address, which is used to set up the sender-to-receiver connection. Classic IP and ARP over ATM is defined in RFC 1577 issued by the IETF (Internet Engineering Task Force). This specification addresses the transportation of IP over ATM, using AAL5 encapsulation for IP and ATM ARP. RFC 1577 introduces the concept of a Logical IP Subnetwork or LIS, where the LIS members are those devices attached to the ATM network. Network members outside of the LIS are accessed via a router. The IP address resolution depends on whether a SVC or PVC is used. In the case of a SVC, the IP address is mapped to an ATM address; with a PVC, the IP address is mapped to a Virtual Channel. It must be understood that it is possible to run IP over ATM networks using the LAN Emulation service. Since, LAN Emulation hides the ATM network from the LAN clients, allowing them to continue using the existing applications and protocols, IP networks can run over ATM using their current IP implementation. LANE delibaretly hides ATM so any network layer protocol that operates over ATM can not gain access to the QoS properties of ATM and must, therefore, use UBR or ABR connections only. In the future, however, this situation is unlikely to endure. In the first instance, as ATM networks proliferate, it is likely that demand will grow to utilize their QoS benefits, since this is one of ATM' s major selling points. In the specific case of IP, the IETF has developed the notion of an Integrated Services Internet. This envisages a set of enhancements to IP to allow it to support integrated or multimedia services. These enhancements include traffic management mechanisms that closely match the traffic management mechanisms of ATM. For instance, protocols such as the Resource Reservation Protocol (RSVP) are being defined to allow for resource reservation across an IP network, much as ATM signalling allows this within ATM networks. RSVP is fundamentlly built upon a multicast paradigm, and routes traffic flows along source rooted point-to-multipoint paths. New multicast protocols like Protocol Independent Multicast (PIM), and their associated unicast packet routing protocols, will hence be closely coupled with RSVP, much as VC routing protocols are closely coupled with UNI and NNI signalling. The IETF is also in the process of developing a new transport protocol, the Real-Time Transport (RTP). RTP is designed to provide end-to-end network transport functions for applications transmitting real-time data, such as audio, video or simulation data, over multicast or unicast network services, and upon transport technologies like ATM for QoS guarantees. The IETF worked first on defining a method for transporting multiple types of network or link layer packets across an ATM (AAL5) connection and also for multiplexing multiple packet types on the same connection. In order to operate IP over ATM, a mechanism must be used to resolve IP addresses to their corresponding ATM addresses. The address resolution table could be configured manually, but this is not a very scalable solution. The IP-Over-ATM working group has defined a protocol to XUlsupport automatic address resolution of IP addresses in RFC 1577. This protocol is known as“classical IP over ATM”and introduces the notion of a Logical IP Subnet (LIS). Like a normal IP subnet, a LIS consists of a group of IP nodes that connect to a single ATM network and belong to the same IP subnet. The operation of the classical model is very simple. It does, however, suffer from a number of limitations. One of these limitations is that the protocol does not attempt to change the IP host requirement that any packet for a destination outside the source node's IP subnet must be sent to a default router. This requirement, however, is not a good fit to the operation of IP over ATM, and a whole class of other“non-broadcast multi-access”(NBMA) networks, such as FR or X.25. In all such networks, it is possible to define multiple LISs, and the network itself could support direct connections between two hosts on two different LISs. The IETF' s“Routing over Large Clouds”(ROLC) working group has been working on protocols that overcome the limitation that noted above. The group is now finalizing work on a protocol known as the Next Hop Resolution Protocol (NHRP). NHRP builds upon the Classical IP model, substituting for the concept of a LIS the notion of a logical NBMA network - that is, a network technology, such as ATM, FR, or X.25, which permits multiple devices to be attached to the same network, but which does not easily permit the use of broadcast mechanisms, as are common on LANs. NHRP will likely be deployed on routers, for use within FR and X.25 networks, amongst others, and it is also likely to be used for router to router communication within some ATM network. Some specific enhancements may need to be made to NHRP. For example, has no support for autoconfiguration. It also today has no support for multicast/broadcast operation; these particular problems whitin ATM networks. Notwithstanding these potential limitations, it is likely that NHRP will play an important role within ATM networks, particulary within the context of the Multiprotocol over ATM (MOA) work currently being done at the ATM Forum. This work will likely involve extending NHRP so as to make it more complete and ATM spesific. Today, there is no specific support in the classical IP protocol for multicast operation. This is a weakness of RFC 1577, particulary in comparison to LANE. Some work has been done to define a mechanism for multicast in RFC 1577. This work attempts to support the IP multicast behavior described in RFC 1 1 12, by a combination of multicast servers and overlaid point-to-multipoint connections. This work may also serve as a model for multicast support in other protocols, possibly including NHRP and MPOA. MPOA Some ATM implementors object to LANE because it is based on bridging, and their networks are routed, and they prefer to keep in that way. Fortunately, there is a solution to bridging vs. Routing dilemma, MPOA. MPOA can be considered an evolution of LANE, which operates by bridging at layer two, because MPOA operates at both layer two for bridging and at layer 3 for routing. Since MPOA is a far more ambitious and complex proposal than LANE, it is expected to provide a XIVsolution to LANE's limitations. For example, MPOA provides direct connectivity anywhere in the ATM cloud, while LANE relies on routers to connect subnets. LANE was intended to hide ATM's QoS from the legacy application, while MPOA will partially expose ATM QoS. The MPOA group is working in close cooperation with the LAN Emulation group. For example, MPOA wants to utilize the LAN Emulation Configuration Server, which allow hosts to automatically configure. Since it was developed specifically for LANE, they have asked the LANE group to make the definition more general so it can be adopted to MPOA. They have also asked LANE to allow the multiplexing of multiple ELANs on to a single connection between two bridges. MPOA is coordinating its efforts closely with the IETF as well, since it will leverage work done by the IETF' s ROLC (Routing Over Large Clouds) group, in particular the NHRP (Next Hop Resolution Protocol) and multicast address resolution (MARS), a means of doing IP multicast address resolution over ATM. The guiding concept behind MPOA is the seperation of switching and routing through the use of a route server. Switching, based on dedicated silicon is fast and cheap. The route server is software running on an inexpensive workstation. When an edge switch receives the first packet, it sends an address query to the route server, which returns the necessary routing information to the edge device. Subsequent packets are forwarded along without any further processing. A router is comparison examines every packet it receives, and is thus slower despite being more expensive. xv
Benzer Tezler
- Asenkron transfer modu ve çerçeve aktarma
Asynchronous transfer mode and frame relay
HALİT DÖNMEZ
Yüksek Lisans
Türkçe
1997
Elektrik ve Elektronik Mühendisliğiİstanbul Teknik ÜniversitesiElektronik ve Haberleşme Mühendisliği Ana Bilim Dalı
PROF. DR. GÜNSEL DURUSOY
- ADSL ve geniş bandlı şebekeler
ADSL and broadband networks
MEHMET DURMUŞ
Yüksek Lisans
Türkçe
2002
Elektrik ve Elektronik Mühendisliğiİstanbul Teknik ÜniversitesiElektronik ve Haberleşme Mühendisliği Ana Bilim Dalı
PROF. DR. GÜNSEL DURUSOY
- Bilgisayar ağları ve güvenlik
Computer networks and security
CÜNEYT BERGEL
Yüksek Lisans
Türkçe
2004
Elektrik ve Elektronik MühendisliğiSakarya ÜniversitesiElektrik-Elektronik Mühendisliği Ana Bilim Dalı
DOÇ. DR. HAKKI CEDİMOĞLU
- Genişbandlı şebekelerde hizmet adaptasyon protokolleri
Başlık çevirisi yok
RECEP EVREN PALANDUZ
Yüksek Lisans
Türkçe
1999
Elektrik ve Elektronik Mühendisliğiİstanbul Teknik ÜniversitesiElektronik ve Haberleşme Mühendisliği Ana Bilim Dalı
PROF. DR. GÜNSEL DURUSOY