Product of the Day: Eternal Systems FTORB Version 2.0 - IONA Orbix/E 2.

by Vendor Written

Product: Eternal Systems FTORB Version 2.0 - IONA Orbix/E 2.Manufacturer: Eternal Systems, Inc.Address: 5290 Overpass Road, Building D                 Santa Barbara, CA 93111Telephone: 805-696-9051URL: www.eternal-systems.comE-mail: sales@eternal-systems.com

Eternal Systems FTORB Version 2.0 - IONA Orbix/E 2.1

Overview

IONA Orbix/E customers can leverage Eternal Systems' FTORB 2.0 to provide application transparent fault-tolerance for their embedded applications. Orbix/E is a lightweight CORBA object request broker (ORB) optimized for embedded applications. Its extreme speed and compactness makes Orbix/E suitable as a software middleware bus in a broad variety of telecommunications, manufacturing, wireless and handheld device applications. Designed for distributed, multi-tiered, embedded computing environments, Eternal Systems' FTORB 2.0 transforms true fault tolerance from a costly, complex, custom programming project requiring specialized knowledge to an affordable, easy-to-use, off-the-shelf solution any programmer can use to make an application fault tolerant quickly.

The combined IONA/Eternal Systems solution provides customers with an off-the-shelf solution for their hard-to-implement, fault tolerance needs, resulting in accelerated time-to-market, significant reduction in development cost, and lower risk in application development

The FTORB replication-based, application transparent, fault tolerance infrastructure speeds the development of true fault tolerance in new and existing applications. FTORB allows application developers to focus on the details of their application programs, using their domain specific knowledge and expertise, without having to worry about the difficult issue of fault tolerance. FTORB requires no specialized expertise or knowledge in fault tolerance programming, eliminates the cost and complexity of custom coding, and works in conjunction with standard CORBA ORB middleware to make true fault tolerance practical for a much wider range of applications than was possible previously.

With the introduction of FTORB, Eternal Systems has transformed true fault tolerance from a high-cost, complex, custom programming project to an inexpensive, easy-to-use, commercial-off-the-shelf collection of software components that can be added quickly to an application.

Continuous Availability

Continuous application availability ensures that - even if a fault occurs in the system -- no data or processing is lost, all application state is retained, and application service continues uninterrupted for the users of the application in spite of the fault.

Eternal Systems' FTORB uses interception-based, active replication to provide true fault tolerance and, thus, continuous availability for the application at the same cost and ease-of-implementation as typical high availability solutions. FTORB is complementary to, and can be used with, hardware-based high availability and fault tolerance solutions. Using FTORB, designers can architect their CORBA-based embedded systems using commercial-off-the-shelf (COTS) components.

Eternal Systems FTORB provides these basic features:

  • Continuous application availability

  • Application transparency

  • Protection of both data and processing against faults

  • Use in distributed, multi-tiered environments

  • Compliance with the CORBA middleware standard

  • Automatic recovery and state management

  • Management of replicas and object groups

How FTORB Works in a CORBA Environment

FTORB Version 2.0 is a small footprint solution that includes etEngine, etGateway, etConsole and thinFT components to enable developers to provide application-transparent fault tolerance for their embedded CORBA applications. No costly application-specific fault tolerance programming is required to implement application-transparent fault tolerance with FTORB.

Located between Orbix/E and the operating system, etEngine is transparent to both the ORB and the application. FTORB is especially effective in environments that require a lightweight, efficient fault tolerance solution where memory and bandwidth are stringent.

Replication-Based Fault Tolerance

FTORB supports a range of replication styles, though most developers choose the active replication style to maintain continuous availability of the application in the event of a fault, by masking the failure of a single replica. In the active replication style, every server replica executes every operation invoked on the replicated server, which ensures that each such server replica is always in the "then current" state.

In CORBA-based applications, client and server objects typically communicate using the TCP/IP-based Internet Inter-ORB Protocol (IIOP). By transparently capturing an application's IIOP messages and diverting them invisibly to a reliable, totally ordered multicast protocol, the etEngine delivers every invocation to every server replica and, thus, facilitates strong replica consistency.

Replication-based fault tolerance, by its very nature, can lead to duplicate operations. etEngine ensures that duplicate invocations are never delivered to, or executed by, the replicas. etEngine also ensures that duplicate responses from a replicated server are never returned to client objects. To conserve network bandwidth, etEngine detects and suppresses duplicate invocations and responses, ensuring that unnecessary traffic is not transmitted over the network.

Managing Replicas and Object Groups

In the FTORB environment, the application administrator can use etConsole to create and manage replicated objects (object groups) and also the replicas that comprise those object groups. Both client and server objects can be replicated and, thus, etConsole allows the user to create both client and server object groups.

Using etConsole, the application administrator can define the location of the replicas on different processors. etConsole also allows the user to query and set fault tolerance properties, such as the replication style, number of replicas, etc. In addition, etConsole supports monitoring of resource usage, processor loads, etc. etConsole is programmed in Java to facilitate application-specific extensions by developers.

Managing Replica State

In an FTORB CORBA environment, an object has three kinds of state: application state (known to and programmed into the object by the application programmer), ORB state (maintained by the ORB for the object) and infrastructure state (invisible to the application programmer and maintained for the object by etEngine). etEngine ensures that all of the replicas of an object are consistent in application, ORB and infrastructure state.

To transfer the application state from an existing replica to a new replica, the application program must be augmented with the get_state() and set_state() methods. The get_state() method retrieves the state from an existing replica, and the set_state() method assigns the state to a new or recovering replica.

Recovery

To provide transparent recovery from faults, the recovery mechanisms of etEngine automatically synchronize all activity among the replicas in the object group, taking the burden off the application programmer. In particular, the etEngine automatically invokes the get_state() and set_state() methods of an application object to transfer the state of an existing replica to a new replica. If, during the state transfer, new invocation and response messages arrive, those messages are queued at the replica while the replica's state is being assigned and then are delivered to the replica. Each etEngine instance on a processor is responsible for storing the invocations, responses and state of the replicas hosted on that processor.

Access to a Fault Tolerance Domain

FTORB employs thinFT and etGateway to provide fault-tolerant, naïve client access to replicated servers within a fault tolerance domain. By using the small footprint, unobtrusive thinFT software, clients can take advantage of the fault tolerance that the replicated servers provide. thinFT communicates with etGateway to gain access to a replicated server within a fault-tolerance domain. etGateway delivers the client's invocations to the replicated server and ensures that the client receives only a single response to its invocation. To ensure that a etGateway is not a single point of failure, that etGateway is also replicated. If an etGateway fails, thinFT clients transparently fail over to another etGateway that provides access to the fault tolerance domain.

An additional benefit of the thinFT/etGateway combination is that client connections into the fault tolerance domain are load balanced, by distributing large numbers of client requests over multiple etGateways.

Availability and Pricing

Eternal Systems FTORB 2.0 running with the IONA Orbix/E 2.1 is available now. Eternal Systems software is available now for various embedded environments, including Linux and VxWorks.

Summary

Developers of embedded systems are experiencing increasing demands to provide continuous availability for a growing number of applications where failure would be extraordinarily costly or catastrophic in economic or human terms. At the same time, they face increased competition and a more highly price sensitive market. Implementing true fault tolerance has been costly and time consuming, requiring specialized knowledge and programming skills, but this is no longer the case.

Eternal FTORB is an inexpensive, easy-to-use, commercial-off-the-shelf set of components that can be added quickly to a CORBA application to provide transparent, scalable fault tolerance. The software's low cost and ease of use eliminates the need for programmers to build custom fault tolerance into their applications, and makes true fault tolerance practical for a much broader range of applications than was possible previously. Combining Eternal Systems FTORB fault tolerance solution with IONA's Orbix/E provides customers with an easier way to develop and write applications, enabling them to integrate applications more quickly.

About Eternal Systems

Eternal Systems software makes it fast and affordable to create true fault-tolerant embedded applications in systems where failure would be extraordinarily costly or catastrophic. Designed for the distributed, multi-tiered computing environments typically found in the data/telecommunications, military, aerospace, industrial control, transportation and medical industries, Eternal Systems' standards-based fault-tolerance infrastructure software enables developers to achieve true fault tolerance in new and existing applications without specialized programming expertise and with the cost or complexity of custom coding.

Founded in 1998 by a team of individuals with more than a quarter century of experience in developing and refining fault tolerance and high availability technology, Eternal Systems is recognized for its leadership role in establishing the Object Management Group's Fault Tolerant CORBA standard. The Company received funding from the California Technology Investment Program, the US Department of Commerce Advanced Technology Program and Smart Technology Ventures. The Company launched its first commercial product, FTORB, in May 2001. The Company has its headquarters in Santa Barbara, CA, and maintains its ties with the Department of Electrical and Computer Engineering at the University of California, Santa Barbara.

Copyright ©2002 Eternal Systems Inc. All rights reserved. Eternal Systems, FTORB, etEngine, etGateway, etConsole and thinFT are all trademarks of Eternal Systems, Inc. All other products and services mentioned are trademarks of their respective companies.

email: potd@ssc.com

Load Disqus comments

Storix backup and disaster recovery