Firma şebekelerinde ATM

Abstract

Firma şebekeleri yeni bir kavram olarak, firmaların kendi özel şebekelerini kamu şebekesi bağlantıları ile desteklemeleri sonucunda ortaya çıkmıştır. Günümüzün hızla değişen rekabet ortamında faaliyet gösteren firmalar, ihtiyaç duydukları bilgi birikimini yerel ve geniş alanda etkili bir biçimde işlemek ve taşımak zorundadır. Bu bilgi birikimi bilgisayar teknolojisindeki gelişmelere bağlı olarak hızla çoğalırken, bilginin taşındığı birimler arasındaki mesafe de giderek artmaktadır. Firmaların kapasite ve boyut olarak ulaştıkları bu nokta, bilgi teknolojisi kullanıcılarının işini oldukça zorlaştırmaktadır. Firma çapında dağılmış farklı tipteki cihazlara, teknolojilere, topolojilere ve protokollere sahip çalışma gruplarını desteklemek, değişen ihtiyaçlara uyum sağlayabilmek ve firma genelindeki bilgi alışverişi için yerel ve geniş alan şebekelerini bütünleştirmek, bu konuda karşılaşılan problemlerin en önemlileridir. Bu problemlere yönelik olarak sunulan çözümlerin mali yönden dengeli olması ve kısa sürede teknolojik gelişmelere yenik düşmemesi gerekmektedir. Firma şebekelerinin gerek özel, gerekse kamu altşebekelerine yönelik olarak çok çeşitli teknolojiler sunulmuş ve sunulmaya devam edilmektedir. Geleneksel şebeke faaliyetlerinde uzun yıllardan beri kullanılmakta olan LAN ve WAN teknolojileri, artan ihtiyaçlara cevap verebilmek için performanslarının üst sınırlarında çalıştırılmaktadır. Diğer yandan, giderek artan şebeke yükünü azaltmak amacıyla şebekelerde değişik düzenlemelere gidilmektedir. Örneğin yerel alanda, şebeke kaynaklarından daha fazla faydalanılması amacıyla şebekenin parçalara ayrılması yoluna gidilmiştir. Buna paralel olarak, tek bir şebeke parçasında yer alan kullanıcı sayısı giderek azalmış ve tek kullanıcılı tahsisli LAN' lara ulaşılmıştır. Geniş alanda ise daha yüksek kapasiteli iletim hatları kullanılmaya başlamıştır. Ancak mevcut teknolojilerle ulaşılan bu çözümlerin maliyetleri giderek artmakta ve istenilen performansı sağlama konusunda yetersiz kalmaktadır. Diğer yandan geleceğin şebeke ihtiyaçlarına çözüm oluşturacak yeni nesil şebeke teknolojilerinin ilk uygulamaları ortaya çıkmaya başlamıştır. Bunlar arasında Asenkron Transfer Modu (ATM) teknolojisi, hem yerel hem de geniş alan şebekeleri için sunduğu avantajlarla dikkat çekmektedir. ATM standartları halen oluşturulma aşamasında olmasına rağmen, daha şimdiden kullanım alanı bulmaya başlamış ve başarılı sonuçlar alınmıştır. ATM teknolojisi aynı zamanda geleceğin şebeke yapısı olan Genişbantlı Tümleşik Hizmetler Sayısal Şebekesi (BISDN) için seçilen iletim teknolojisi olup, geleceğin şebekeler arası bağlantı lan için en önemli çözüm olarak görülmektedir.

Enterprise networking, refer to corporate networks that span multiple company departments, such as sales, marketing, engineering, manufacturing etc. Enterprise networks include both data processing networks and departmental end-user networks. Initially, many of these networks were standalone islands and it was difficult to transfer data between networks. Now the trends is toward interconnection and consolidation. Enterprise networking is based on the concept of internetworking and interoperability of computer systems and networking products. The concept allows the exchange of data, applications and network services between dissimilar computing platforms and local area networking environments via the enterprise network. It also facilitates the management of the enterprise network from a centralized point. Evolution in computing and communications technology has enabled the rapid exchange of business information on a global basis. Business information includes customer orders, product inventories, accounts, revenue and profit. When wide area business information was voice telephone traffic, telephone companies were the primary supplier of the networks for carrying this traffic. Rapid advances in customer applications technology have forced these companies to develop, deploy and operate a public network according to users' information needs. This difference between users' needs and carrier capability has created a market for customer premises equipment-based private networks. To create these networks, users combine equipment from manufacturers with public transport or switched services to create an enterprise network. This network satisfies their inter-enterprise and intra-enterprise information movement. Enterprise networks allow the sharing of information among the various departments and organizations of a corporation. It is not important whether the locations are in the same building or across the globe. These networks uses the considerable computing power of the corporation for improved productivity. Operationally there is a split of enterprise networks into premises networks and public network services. This split allows to optimize the networking technology for individual user requirements. In designing such networks it is important to create an enterprise network that is ideally protocol independent and scaleable. That provides an infrastructure for systems interoperability in a multi-platform, multi-vendor environment. The enterprise network can also be viewed as a super VIIbackbone network which can be used to connect local backbone networks across the organization. An enterprise network must serve the voice and data needs of the company. It would be ideal, if the network could provide end-users with services like multimedia, videoconferencing, electronic mail, FAX, etc. Enterprise networking is derived from classical private and public network architectures. It does not interest with who owns or controls the components of the network. Instead it includes media, customer premises equipment, public services, and information management that is required to satisfy the information management of a corporation. Today, a big change is going on in the information processing environment. This change is affecting any business which uses information. For many corporations, adaptation is very difficult and the result could be a chaotic environment. At an elementary level, the change is the move away from centralized, mainframe-based information processing toward desktop processing, done by the departments which will use it, and performed where needed. The departments or functions in a business are all linked, and at some point in time, at some level all business units need to exchange information, combine it or work together. The need to move information among these departments, groups, or geographic sites, implies a communication infrastructure that can accomplish the task reliably, quickly and at affordable cost. There are several proposals for accomplishing information movement, each of uses different components. Centralized, mainframe-based information processing is relatively expensive and complex. Mainframes has to be run 24 hours a day, attended by a large, highly-trained staff, and used under complex, tightly-controlled, centralized procedures. This insures the integrity of the information being processed. Most mainframe users relies on one key vendor. By contrast, personal computers (PCs) can be placed in the departments where they are actually needed and used. They are easier to use than mainframes, more flexible, and require less training. Data integrity and reliability can be maintained more easily, with only a small training effort. Communication needs are constantly changing. Enterprise networking will become the backbone of any successful company in today's evolving information environment. Enterprise networks will be a company's most important part that deliver vital information anywhere in the world and to everyone that the company does business with, like employees, suppliers, customers etc. The problems facing users, focus on supporting enterprise-wide work groups on dissimilar computing equipment and network devices operating on multiple protocols, different network topologies, changing user needs, and the need to integrate local and wide area networks to communicate over the enterprise.Users are seeking cost- effective and long-term solutions to these problems that will not be replaced by a better solution within a short time. VIIIInitially, there was no desktop-to-desktop communication and PCs were not linked to each other. Communication with mainframes was possible via modem or via special communication hardware. The need to communicate easily between desktops gave birth to Local Area Networks (LANs), Wide Area Networks (WANs), bridges, routes, hubs, and others. These networking technologies now compete for the attention of the executives, responsible for information technology. Central office switch makers also competing for this market, and trying to get onto customer's premises. Another faction is the telecommunications carriers. They are trying to convince customers that, their experience, their size, and their offerings mean they are the most appropriate vendor to manage customer networks. Other competing forces are computer companies who are processor/server vendors and they have strong solutions for client-servers and distributed architectures. Competition occurs across these specific technologies. Vendors from the computing side are competing with vendors from telecommunications side. Router vendors are in competition with hub vendors and with WAN vendors. WAN vendors compete with router vendors, with hub vendors and with carriers who are offering to manage customers private networks. This competition presents a confusing and risky situation for the market seeking to plan or maintain networks. The most common problem in networking performance is bandwidth. Rising bandwidth requirements have exceed the capacity of existing LAN and WAN technologies. R is necessary to use a new form of transport and switching technology to satisfy current and future bandwidth requirements. This technology must be scaleable of multi-megabit to gigabit data rates and it must support any form of information medium. And it must act as a common transport for both LANs and WANs. This new technology, called Asynchronous Transfer Mode (ATM) is proposed as the solution for internetworking requirements into the future. ATM is a form of fast packet switching, employs short, fixed length packets called cells and is commonly referred to as "cell relaf. The actual cell size is 53 bytes in length. The first five bytes form the cell header while the remaining 48 bytes form the information field or payload. The fields within the cell header contain several information fields to insure that the packet is transmitted to its specified destination. The header fields include information to regulate traffic flow between the customer premises equipment and an ATM switch to prevent overloads and congestion. They also include the address of channel sending the information and its routing address which defines a logical path through the network, and finally includes an error control field which protects the header from bit errors. Cells are transported over network links called virtual circuits. Multiple cells from different sources share the same link through the network from origin to destination. There are two types of virtual circuits: Permanent virtual circuits (PVCs) ani switched virtual circuits (SVCs). A PVC behaves like a dedicated line and when activated it will always establish a path between source and destination end points. SVC is like public switched IXtelephone service in that calls can be made dynamically between a source and destination end point in the network. There are certain benefits to use ATM technology: ATM can transport and switch both packet oriented (LAN) and isochronous (voice, video) traffic. The fixed cell format and small cell size of ATM are ideal characteristics for predictable throughput and short delay, also called latency. ATM supports voice and packet data transmission for use in a multimedia workstation. The ability to transport and switch both isochronous and packet data leads to support of future multimedia applications. The small fixed-sized cell can be transported through a network faster than larger, variable-length frames or packets. A fixed-size cell can be processed by faster switching algorithms than those for variable-length packets. The key attribute of ATM technology is its ability to dynamically allocate bandwidth on demand. ATM represents a shift for LANs from multiple-access, shared-media to switched, point-to- point connections. Star-configured ATM hubs and switches simplifies physical reconfiguration. Since ATM uses switched point-to-point connections and stations can be interconnected at the full bandwidth of the media, ATM eliminates the need for LAN segmentation. Since ATM utilizes point-to-point transmission via a physically switched connection instead of shared-media communication, ATM provides much more security. The migration from existing technologies to ATM is a major user consideration. Users must consider changes to the existing infrastructure and they must be sensitive to the disruption and cost of migration. The transition from existing technology to ATM must be gradual. ATM can utilize today's existing cabling infrastructure and consequently existing networked applications can run over ATM unchanged. ATM network interface cards will contain a simple adaptation function that will make the ATM network transparent to the user's software applications. ATM has been defined in the context of WAN standards for B1SDN. However, the first deployment of ATM will be in local area networks and in the customer premises. The first ATM products to be available to the users are from LAN internetworking vendors in the way of hubs, switches, and routers. ATM switch architecture is designed for low latency and can achieve significant statistical multiplexing gain for bursty LAN traffic. Therefore, ATM switches which interconnecting departmental LANs can provide a much higher throughput. ATM switches are already available from many vendors. ATM switch based backbone LANs are the first commercial deployment of ATM. ATM bridges/routers are required to bridge or route, and multiplex the traffic from conventional departmental LANs (Ethernet, Token Ring, and FDDI) to an ATM backbone LAN. These devices are interconnected by ATM switches that comprise the backbone LAN infrastructure. The ATMbridge/router exists as a standalone device or as a plug-in module to an intelligent wiring hub. ATM hubs concentrate low-speed virtual connections, such as from individual LAN workstations to a high speed virtual connection. The hubs high speed virtual connection will connect to an ATM switch. Some ATM hubs also perform local ATM switching. ATM hubs perform adaptation to convert between conventional LAN packets and ATM units. ATM hubs can also include an ATM bridge/router module to bridge or route the converted packets onto an ATM backbone. The next phase of the ATM evolution incorporate a local ATM switch fabric within a hub. This architecture will be used for the workstations, servers, and terminals with direct ATM interfaces. The next evolutionary phase for LAN applications will be the deployment of multimedia applications. ATM is the best technology for transporting multimedia. Multimedia applications are not in widespread use, but one of the significant features of ATM is that, it allows a seamless migration to multimedia applications. These applications accelerate the use of ATM LANs and a multimedia application on an existing ATM LAN can be easily accommodated by simply adding a new adaptation layer capability to the end station. No changes are required to the ATM switch. The presence of ATM in the wide area networks will be initially in the form of connectionless public network service which is both a metropolitan area and a wide area service offered by the major carriers. In addition, several carriers begin to the ATM trials primarily with PVC services. After the standards are available, SVC services will also be offered. Private backbone networks using proprietary cell relay WAN switching nodes are primarily being used to transport data across a WAN, and are not used for delay-sensitive isochronous traffic such as voice. Voice will continue to be transported in both private and public networks using traditional circuit switching until BISDN emerges late in the decade. Early BISDN deployments will provide access to the public network via a hybrid mapping of both isochronous voice and ATM cells. Since the development of standards on BISDN began on 1986, enough of the ATM technology has been defined so that it can now be deployed in customer premises networks and private PVC WAN environments. Public BISDN service will not become available to interconnect private ATM networks in the immediate future. Firstly, BISDN using PVCs will become available in major metropolitan areas only. After then BISDN using SVCs will begin as descrete islands of service, (regional BISDN) These islands will then be interconnected to reach the global ISDN.