【正文】 轉(zhuǎn)變的背。 J2EE編程模型一直被認(rèn)為是分布式的編程模型，在該模型中應(yīng)用組件在 J2EE 服務(wù)器上運(yùn)行并且彼此可以相互交互。 企業(yè)信息系統(tǒng)（ EIS）或數(shù)據(jù)層 這一層指的就是企業(yè)信息系統(tǒng)，比如關(guān)系數(shù)據(jù)庫(kù)， ERP 系統(tǒng)，消息系統(tǒng)等。同步調(diào)用的 EJB 組件通過(guò)通常被 EJB 部署者綁定在 JNDI中的工廠代理對(duì)象來(lái)表現(xiàn)自己。 EJB 規(guī)范定義了幾種類型的組件。這些組件提供了持久化機(jī)制和事務(wù)支持。表現(xiàn)層的服務(wù)器端通常通過(guò) Http（ s）協(xié)議來(lái)進(jìn)行訪問(wèn)。 這些組件負(fù)責(zé)把業(yè)務(wù)數(shù)據(jù)傳遞給終端用戶?？蛻舳税g覽器， applets， Java 應(yīng)用程序等和負(fù)責(zé)與服務(wù)器端的表現(xiàn)層或者業(yè)務(wù)層進(jìn)行交互。屬于各層的 J2EE 組件在開(kāi)發(fā)時(shí)遵守具體的 J2EE 標(biāo)準(zhǔn)。 在眾多的服務(wù)列表中，應(yīng)用服務(wù)器必須提供消息通信，事務(wù)處理，命名機(jī)制和其它應(yīng)用組件用到的服務(wù)。 (4) 各種各樣的人物角色。 (2) 組件和主服務(wù)器的鏈接。 J2EE 是開(kāi)發(fā)多層企業(yè)應(yīng)用 JAVA程序的綜合性的標(biāo)準(zhǔn)。 圖 1 J2EE的三層結(jié)構(gòu) 組件框架為組件的執(zhí)行提供了一個(gè)集成的環(huán)境，從而顯著的減少了在設(shè)計(jì)，實(shí)現(xiàn)，部署和維護(hù)應(yīng)用程序時(shí)工作。應(yīng)用組件“插入” 確立它們運(yùn)行環(huán)境和規(guī)定它們交互的組件框架中。 2 J2EE背景知識(shí) 介紹 組件框架。第四部分更深入的描述了有關(guān)這種架構(gòu)特別重要的和有趣的內(nèi)部機(jī)制。本文剩余部分結(jié)構(gòu)如下：第二部分提供了必要的背景以理解和研究有關(guān)的 J2EE組件 技術(shù)規(guī)范。 JBoss的組件結(jié)構(gòu)允許根據(jù)部署應(yīng)用程序的需要增加服務(wù)配置。我們把這種架構(gòu)作為 JBoss開(kāi)源 java應(yīng)用服務(wù)器的一部分加以實(shí)現(xiàn)，在幾個(gè) J2EE樣本程序比如 Java PetStore,， RUB和 TPC_W_NYU中進(jìn)行測(cè)試。組件配置過(guò)程評(píng)估了應(yīng)用程序路徑的正確性，確認(rèn)在系統(tǒng)組件上的應(yīng)用組件的獨(dú)立性和完成復(fù)制組件的部署。一種為配置組件屬性而開(kāi)發(fā)的定義和其表述性語(yǔ)言允許內(nèi)部組件間獨(dú)立的規(guī)范和組件間屬性的繼承。它使得應(yīng)用組件與系統(tǒng)組件中清晰的分開(kāi)。這種框架為組件定義了結(jié)構(gòu)描述語(yǔ)言（ ADL），鏈接說(shuō)明和集合。 (5) 提供上述便利而不會(huì)增加應(yīng)用程序員的設(shè)計(jì)負(fù)擔(dān)。 (3) 提供簡(jiǎn)單但可表達(dá)的抽象方法去控制通過(guò)部署和拆卸組件獲得的適用性。 這種分布 式配置框架必須滿足： (1) 聲明內(nèi)部組件的一致性規(guī)范和定義其對(duì)組件配置部署的影響。例如你必須在配置和部署應(yīng)用組件前先建立JDBC數(shù)據(jù)源、設(shè)立消息目的地及資源適配器。然而，為了實(shí)現(xiàn)這種動(dòng)態(tài)可適性，我們需要一種框架來(lái)在這樣的環(huán)境里自動(dòng)化地配置 J2EE 應(yīng)用程序。 (5) 新的請(qǐng)求路徑可以復(fù)用先前的組件配置路徑。 (3) 新加的復(fù)制組件可以動(dòng)態(tài)配置以滿足新的請(qǐng)求。其研 究的主要結(jié)論可以概括如下： (1) 如果應(yīng)用合適的應(yīng)用程序，則在廣域網(wǎng)中的垂直負(fù)荷可以察覺(jué)到延遲。垂直分割（例如運(yùn)行網(wǎng)絡(luò)層和事務(wù)層在不同的虛擬機(jī)）被用于錯(cuò)誤分離和均衡負(fù)荷，但是由于它遠(yuǎn)程調(diào)運(yùn)的大量使用顯著地增加了運(yùn)行時(shí)時(shí)間導(dǎo)致其實(shí)現(xiàn)很不實(shí)際。這種應(yīng)用程序通常被部署在公司的內(nèi)部網(wǎng)或者是因特網(wǎng)上，以高事務(wù)容量，大量的用戶和覆蓋范圍廣的訪問(wèn)為特征，它通常會(huì)被部署在中央?yún)^(qū)域，采用服務(wù)器集群來(lái)均衡負(fù)載（平均分配）支持用戶下載。這種需要對(duì)于哪怕在單一的應(yīng)用程序服務(wù)器上嘗試部署 J2EE應(yīng)用的人來(lái)說(shuō)也很明顯，這種任務(wù)涉及到大量的系統(tǒng)服務(wù)和應(yīng)用組件的配置。 1 Introduction In recent years, we have seen a significant growth in ponentbased enterprise application development. These applications are typically deployed on pany Intras or on the Inter and are characterized by high transaction volume, large numbers of users and wide area access. Traditionally they are deployed in a central location, using server clustering with load balancing (horizontal partitioning) to sustain user load. However, horizontal partitioning has been shown very efficient only in reducing applicationrelated overheads of userperceived response times, without having much effect on workinduced latencies. Vertical partitioning (., running web tier and business tier in separate VMs) has been used for fault isolation and load balancing but it is sometimes impractical due to significant runtime overheads (even if one would keep the tiers on a fast localarea work) related to heavy use of remote invocations. Recent work [14] in the context of J2EE ponent based applications has shown viability of vertical partitioning in widearea works without incurring the aforementioned overheads. The key conclusions from that study can be summarized as follows: ? Using properly designed applications, vertical distribution across widearea works improves userperceived latencies. ? Widearea vertical layering requires replication of application ponents and maintaining consistency between replicas. ? Additional replicas may be deployed dynamically to handle new requests. ? Different replicas may, in fact, be different implementations of the same ponent based on usage (readonly, readwrite). ? New request paths may reuse ponents from previously deployed paths. Applying intelligent monitoring [6] and AI planning [2, 12] techniques in conjunction with the conclusions of that study, we see a potential for dynamic adaptation in industrystandard J2EE ponentbased applications in wide area works Through deployment of additional application ponents dynamically based on active monitoring. However, in order to achieve such dynamic adaptation, we need an infrastructure for automating J2EE application deployment in such an environment. This need is quite evident to anyone who has ever tried deploying a J2EE application even on a single application server, which is a task that involves a great deal of configuration of both the system services and application ponents. For example one has to set up JDBC data sources, messaging destinations and other resource adapters before application ponents can be configured and deployed. In a wide area deployment that spans multiple server nodes, this proves even more plex, since more system services that facilitate internode munications need to be configured and started and a variety of configuration data, like IP addresses, port numbers, JNDI names and others have to be consistently maintained in various configuration files on multiple nodes. This distributed deployment infrastructure must be able to: ? address interponent connectivity specification and define its effects on ponent configuration and deployment, ? address application ponent dependencies on application server services, their configuration and deployment, ? provide simple but expressive abstractions to control adaptation through dynamic deployment and undeployment of ponents, ? enable reuse of services and ponents to maintain efficient use of work nodes’ resources, ? provide these facilities without incurring significant additional design effort on behalf of application programmers. In this paper we propose the infrastructure for automatic dynamic deployment of J2EE applications, which addresses all of the aforementioned issues. The infrastructure defines architecture description languages (ADL) for ponent and link description and assembly. The Component Description Language is used to describe application ponents and links. It provides clear separation of application ponents from system ponents. A flexible type system is used to define patibility of ponent ports and links. A declaration and expression language for configurable ponent properties allows for specification of interponent dependencies and propagation of properties between ponents. The Component (Replica) Assembly Language allows for assembly of replicas of previously defined ponents into application p