The strength of Cast Iron
Comment Appliances seem to be getting everywhere. One area for their use that I have not previously looked at, are appliances for integration. As far as I know, there are three vendors in this space: BridgeWerx, which is based around Microsoft's BizTalk; Cast Iron Systems; and InfoTone.
Now, as regular readers will know, my primary interest is in data and so I am more interested in data integration than application integration. As a corollary, I am not especially concerned with BridgeWerx as it is clearly focused on application integration only (and specific integrations at that).
I am also less interested in InfoTone, since it is implementing its connectors in hardware as a part of the appliance. The problem with this approach, it seems to me, is that this means that connectivity is likely to be limited and not easily extensible.
Thus, my interest is with Cast Iron Systems. The fact that the company already has a partner in the UK and expects to be making a push into Europe next year only serves to reinforce that interest.
Cast Iron was founded in 2001 and started shipping its first product, known as the "Blue Box" in 2003. Just recently, it has introduced a "Red Box". Put simply, the Blue Box is a hardware appliance on which integration software has already been pre-installed and configured, so that you can "plug-in and go". Connectors at the front and back-ends plug in as required.
The Red Box, on the other hand, has a pre-built Salesforce.com connector already installed at the back-end, so it is simply a question of choosing a connector for your source data. There is an option that supports multiple front-end connectors and you can, of course, have multiple boxes and there is a web-based management console that will monitor your environment across multiple boxes.
Cast Iron, like the other vendors mentioned, is focused on application integration, but I am interested in its wider applicability. In other words, how good is it at data integration, either database to database or application to database, or even database to application?
Certainly, the company has users that employ the software for data integration, though this is typically at table level. Well, transformation and mapping is pretty much like a conventional data integration tool. That is, you drag and drop on a palette. There is also a mapper for flat files and a mapper for when you don't have any metadata. There is also scheduling, and change data capture can be used for real-time purposes.
Where there aren't any facilities are for data governance: there are no profiling, matching or cleansing facilities at all and there are also no facilities for doing such things as impact analysis.
This is arguably less important for application integration and it probably isn't very significant for loading data into Salesforce.com. Moreover, this is not necessarily a deal breaker anyway: there are some elements of profiling that you can build into transformations and you could, after all, do these data quality things separately, either by off-loading them or doing them on the source and target. Nevertheless, this is likely to be off-putting for large enterprises looking at a platform-based approach.
However, Cast Iron isn't targeting large enterprises. Although it has companies like British American Tobacco and Toyota as customers, it is primarily targeting the mid-market. Interestingly, it has also changed its pricing model so that it uses the same subscription-based pricing as Salesforce.com. Thus, an entry level Red Box is $3,000 per month including support. Given that there are no implementation costs whatever, this looks pretty attractive.
To conclude, it is clear that Cast Iron's strength is in application integration and it offers comparable performance to EAI solutions but without the implementation pain. For certain types of data integration tasks companies may make use of Cast Iron, but I suspect that this will be where they are already using one of the company's boxes for application integration.
Copyright © 2006, IT-Analysis.com
Sponsored: Benefits from the lessons learned in HPC