How large will Dieselpoint Search scale?
See Dieselpoint Search Software – Scalability
What are the system requirements?
Dieselpoint Search will run on both low-end and high-end hardware. Current customers are using our search software on both low-end Intel servers and high-end Sun and IBM equipment.
Our search software is written entirely in Java, and is truly cross platform. It’s currently in production use on Intel/Windows, Intel/Linux, Sun Sparc/Solaris, IBM/iSeries/AS400, and Mac/OS X.
It will run in 128 mb of RAM or less, although more memory will yield better performance. The system caches as much of the index as possible in memory.
It ships with an internal webserver, which is adequate for most applications. For more involved applications it is common to use a commercial J2EE server. We currently have customers using Websphere, BEA Weblogic, Resin, and Tomcat.
Can I run it off a CD? Can I distribute it to run on laptops?
Yes, and yes, subject to license restrictions. We have customers doing both successfully. You’ll need to contact us for licensing terms.
Can I do hit highlighting?
Yes. You can also show document summaries, with hits highlighted in context, in the typical search-engine style.
Can I create custom views? I want only certain people to see certain subsets of my data.
It’s not difficult to do this. For example, suppose you have a product catalog and you want retailers to see some products and end-users to see others. Add a code to each item for “retailer” or “enduser”. The system supports multi-valued attributes, so you could add both codes to the same item. Then, when a user runs a query, you append a filter depending on who the user is. If the user is a retailer and he or she searches for “widgets”, you modify the query to be “widgets AND (code=retailer)”.
How do I update an item already in the index?
An update is implemented as a “delete” plus an “add”. The simplest way to handle updates is to make sure that each item has a unique id. When adding a new item the system will search for an existing item with the same id, and if it exists it will be deleted.
What types of documents and data can Dieselpoint search?
The system supports a wide variety of data formats, including MS Office, XML, PDF, flat files, and SQL databases. It will support any arbitrary format if you write directly to the API.
Is there any preferred file format for catalogs?
Not really. The system can extract data from SQL databases, XML, or other structured sources.
XML is a good choice for many applications. The system supports XML indexing in any format. It provides added support for XML in the Dublin Core and ECCMA formats. Building product catalogs with data in the ECCMA format is particularly easy.
Does it require a database?
No, data is stored in our own file format. Databases aren’t fast enough to handle heavy throughput.