Web application server configuration and parameter settings usually require a large amount of important and highly inter-dependent parameters, making it very difficult to design an appropriate Web application server setup. This is why Cosminexus provides several typical system configuration patterns. Since related parameters are grouped together, only a minimum number of parameters need to be set, so that the appropriate server configuration can be set up instantly.
JP1/IM - CM (JP1/Integrated Management - Central Information Master) functionality can be used to define a Web system configuration. The recommended Web model for JP1/IM - CM (Web server, application server, and load balancer) can be used to pass definition information to Cosminexus simple setup functionality, to perform Cosminexus (Web server, application server, and load balancer) settings.
Applications can now be deployed in their directory, so that modified classes are automatically reloaded. Until now, even switching only one class required processing such as creating an archive, stopping and deleting an application, and importing a new application, but now switching is performed simply by updating the classes.
When an application is changed, its operation needs to be checked in the production environment, which generally involves stopping operation. Cosminexus supports functionality to test applications in the production environment without stopping running operations, to streamline operation.
Cosminexus can be used in coordination with JP1/AJS2 (JP1/Automatic Job Management System 2), a JP1 job management product, to schedule Cosminexus starting and stopping, as well as the starting and stopping of business applications running on Cosminexus. This helps makes operations more automated.
JP1/Integrated Management, a JP1 integrated management product, enables centralized management of operational statuses and events for operations, networks, servers, and applications on various kinds of platforms. It also enables all events on an enterprise information system to be managed visually through simple operation, to reduce management-related burdens. It can also be linked with JP1 to monitor JP1 events issued by Cosminexus, for centralized management of operation and performance monitoring.
Monitoring trees used for monitoring Cosminexus are automatically generated, and can be easily customized through GUI operation. Failure can be detected in the monitoring tree window of JP1/Integrated Management, enabling the Cosminexus operation management portal windows to be called from the monitoring tree window so that troubleshooting operations can be performed in a sequence.
Cosminexus can be used to monitor the usage status for various resources, such as Java VM heap memory areas and threads. Usage statuses can be understood for tuning and proactive failure prevention, and actions can be defined in advance for when thresholds are exceeded, to handle resource degradation automatically.
For example, when a full garbage collection (full GC) predictor is detected based on the usage status of the Java VM heap memory area, the active server and standby server can be switched before full GC occurs. This prevents slowdowns due to full GC, to improve performance of system operation.
JP1/Performance Management, a JP1 availability management product, comes with an Internet service monitoring agent (JP1/Performance Management - Agent Option for Service Response). This agent enables measurement of service performance (service response times) for not only the HTTP protocol, but also various other protocols including FTP, DNS, SMTP, POP3, and IMAP4, enabling the response time of a series of Web operations to be measured in simulation. User commands can also be used to measure and collect response times for arbitrary services. This enables system administrators to use service response times as a guide to instantly check whether the quality of service has degraded.
JP1/Performance Management can be linked with uCosminexus Application Server monitoring agent (JP1/Performance Management - Agent Option for uCosminexus Application Server) to enable real-time monitoring of the operating performance status for a Cosminexus system, enabling the causes of service quality degradation to be discovered quickly.
JP1/Performance Management monitoring agents for operating systems and various applications can be used to link with robust reporting functionality (real-time reports and historical reports), enabling intuitive understanding of the Cosminexus system operation status.
Since Web systems are processed across multiple servers, it is very difficult to determine where an issue is occurring. Cosminexus can use Hitachi's field-tested performance analysis trace (PRF trace) functionality to obtain logs for each functionality point. Since these logs have unique identifiers for each request from clients, they facilitate tracking significantly. This is especially true when HiRDB or Oracle is used as the database, because database connection IDs are also output to the logs, to enable an understanding of the processing sequence down to database access.
Even when failure occurs on a J2EE server, session information can be carried over to another server so that operation can continue. Since operation can continue without reporting errors to the client, failure reporting does not become a nuisance.
Cosminexus can monitor application execution times on the method level. When a database processing delay prevents an SQL command from returning for long durations, or an infinite loop occurs within an application, a timeout can be reported and the method can be cancelled, to localize failure impact and quickly recover system operation.
Until now, preparing for server failure required the setup of a standby server for each server machine. Cosminexus supports N:1 clusters, which enable one general-purpose standby server to be set up for multiple servers, to greatly reduce costs related to ownership, setup, and operation. When failure occurs, the general-purpose standby server can complete transactions being processed, so that operation can continue.
This cluster environment comes with HA monitor, which provides functionality to automatically and quickly perform high-speed hot-standby and other types of switching, to improve system performance and reduce operational costs.
ISO/ICE 15408 is an international standard for evaluating whether information systems and products have been properly designed and implemented from a security standpoint. Using products certified for ISO/IEC 15408 not only enables more secure system setups, but also helps clear the security requirements for various systems. Starting from version 7.0, uCosminexus Application Server has obtained this ISO/ICE 15408 certification and now enables setup of even more secure systems.
Note that this certification does not guarantee the products themselves, but indicates that that the evaluation results meet the assurance requirements of ISO/IEC 15408.
Cosminexus provides an application development and runtime environment that supports Web services conforming to SOAP, WSDL, and other specifications. It enables setup of secure systems, by providing APIs that handle WS-Security-compliant security functionality as standardized by OASIS and XML security compliant with XML signing and XML encryption as standardized by the W3C.
Since request queues for operations awaiting processing can be defined for each application, mission-critical operations can be given priority and concurrently executed thread counts can be defined for each of multiple operations, to control flow and prevent impact on other operations when requests are concentrated on another operation. Since service blockage can also be performed by operation, certain operations can be temporarily stopped without changing Web server settings.
Throughput control can also be controlled by URL, to prevent particular servlets or application logic that involves heavy processing from impacting other operations. This functionality works particularly well with Struts and other frameworks based on an MVC model, by offering detailed flow control.
Cosminexus offers support for a scheduler that inherits online technology tested in the field with OpenTP1, to provide functionality such as the following that ensures stable response times and throughput even under heavy loads.
Load balancing can be performed in two steps: when requests are allocated to the server, and when requests are actually registered in queues. This enables dynamic and efficient load balancing based on the server load information.
Also, since the number of threads used can be defined for each scheduling queue just as for front-end Web systems, operation flow can be controlled, and processing can be given priority for certain operations. Priority control can be controlled for each request, so high priority requests can pass through business systems that hold requests in queues for extended periods.
