Only a few products are available, but Candle Corp is still talking up its move away from its MVS mainframe base and into the world of Unix. The move was first announced back in March (CI No 2,386) and the privately held, Santa Monica, California-based company says the foundation of its move ot Unix is Candle Technologies. IBM Corp’s Automation Center/400, which was jointly developed with Candle, is based on the architecture and has been available since March. But other products, which have already been demonstrated at various shows, are not likely to get to market until next year. Candle Technology is split into four parts: workstation client software for OS/2, Windows and Motif systems; the data server-repository, based on Candle’s own relational and object database and running under Unix, OS/2, Windows NT, OS/400s or MVS; applications services, based on a set of C++ libraries from Borland International Inc, although Candle wrote its own C++ compiler for the mainframe; and communications services, a layer of software to handle communications protocols. Next year a monitoring and management system for BatchPipes/MVS based on Candle Technologies will be available, as will a Unix monitoring system for AIX, HP-UX and Solaris; and a monitoring system for Windows NT 3.5. Candle is also talking to IBM about products for the SP1 PowerParallel system and so-called parallel mainframes and says that by the second quarter of next year monitors will be available for CICSplex and Sysplex environments. Candle adds that over the next year and a half it will bring to market monitoring systems for NetWare, Solaris, distributed databases and Simple Network Management Protocol gateways. Candle describes itself as a case study of a legacy mainframe company trying to make the transition to the new world. As for the new world, Candle describes its has having five levels of computing. The first tier is the management of discrete, single systems and costs are fairly predictable. The second tier is the management of heterogenous systems where there is often no common interface for the various systems in use. This means that costs are increased and can only be controlled if management is from a centralised point: this is where Candle Technologies plays a part by enabling users to direct data to a central server where it is processed. The server is independent of the systems being monitored and managed and the use of probes does way with the need for constant polling of network nodes as the probes send alerts only when problems. Tier three is about managing groups of the mixed systems managed as individual systems in tier two. Candle says this is the cost and technology wall that many companies are hitting. Applications that work fine at the second tier fail at this level. If probes are to be used, then they have to be installed on thousands of nodes and if that was achieved it would take only a few alerts for the network to become burdened. Aimed at this level is Candle’s Visual Programming, designed to enable the creation of alert and automation rules as objects that can then be distributed to nodes using drag and drop. There is also an automated library for these objects that ensures that should any rule be changed it will only be installed on systems where the previous version was running. In Automation Centre/400, automation rules are sent out from the host once and then reside on the desktop. The rules will handle the main issues; anything it cannot deal with there will generate a prompt to ask the user. Tier four is merely a more and more heterogenous system but tier five in Candle’s definition covers managing business applications over these multiple, heterogenous applications.
