When we started developing completely new Feedreader3 about one and half years ago one big functional request was to separate user interface and database engine. This gave us a really nice result – if we subtract Feedreader user interface from Feedreader3 all that is left is a really powerful feed management component that can easily refresh thousands of feeds and offer all the features that Feedreader3 has. We called this component Newsbrain.
We are using Newsbrain in multiple applications. It's included in desktop version of Feedreader. And it's also included in Feedreader Connect Enterprise.
So what exactly is Feedreader Connect Enterprise? Feedreader has always made it possible for users to search and find specific information from big number of articles. For this we developed smartfeeds. But what if you really want to create persistent smartfeed that searches for information from a lot of portals and newspapers. One possibility is to use one of multiple web services that offer this kind of functionality (google, technorati,...). But what if this is not enough and you have specific needs? Then comes the Feedreader Connect Enterprise.
Feedreader Connect Enterprise is basically a special high performance version of Feedreader that runs as Microsoft Windows Service (without user interface) and uses Microsoft SQL Server as backend.
To view the possibilities of Feedreader Connect Enterprise from more practical point let me tell you about real world usage example for this application.
Our company does not only develop Feedreader related products but also offers some services for media related companies and organizations in Estonia. After all we have been in media related business for almost 10 years and know specifics of it.
One service that we provide for our clients is monitoring Russian online media for information and articles about Estonia. At the moment there's a lot of talk about „Bronze Soldier“ in Tallinn, capital of Estonia. It's really sensitive matter for Russia and Russians , so in Russian media they talk about this matter every day. So our mission is to monitor Russian media. At the same time there does not exist any international web services that let us input „estonia“ in Russian language to search box and constantly receive newest articles about us (the Estonians) in Russian media. So only possibility is to use award winning Feedreader to perform the task.
Russia has about 1000 different big online news channels that all produce original content. And to accomplish the task we just subscribed them to Feedreader Connect Enterprise - as easy as in desktop version of Feedreader. But here comes the performance difference between normal Feedreader and Feedreader Connect Enterprise. All those 1000 online channels are really active and are updated frequently. We are not talking about one article per day. It's more like multiple articles per hour per one news channel. Based on this information total traffic is huge to say at least ;). Feedreader Connect currently receives approximately 1100+ new articles per hour. It uses all its 16 threads to poll for updates on servers. This cannot be done by some standard solution (Feedreader, Feeddemon, ....).
But this is not all. As we know not all news feeds contain full text of articles. Usually you have to jump to original page to read complete article text. Sometimes we all have encountered message „article is expired“ on remote web page. This is OK then just reading articles. But in our monitoring service we want to save every piece of information that we get from all those servers. So the natural need is to cache all articles locally. And this is second task that Feedreader Connect Enterprise does – all articles are downloaded to local hard disk.
And what's the result for the hard work that Feedreader does related to our monitoring service? Feedreader receives 25000+ new articles every day. 150 of those articles are about Estonia. This means that you really must do a lot of work to get even few results. But we are really happy with those quality results. :)
Technical background : Our Feedreader Connect Enterprise application runs on HP small office server with P4 processor, 1.5G of memory and two normal hard disks. We are using Windows Small Business Server 2003 as operating system and Microsoft SQL Server Express 2005 as database backend. Server constantly uses 100% of 1Mbit connection that we gave him. :)