Note that there are some explanatory texts on larger screens.

plurals
  1. PO
    text
    copied!<p>Sharing data in the three ways you've out-lined (via Web Services) is indeed possible, but there's lots of other factors which will drive the right approach.</p> <p>Architectural aspects to consider:</p> <ul> <li>Data Architecture (business ownership, security, management and so on).</li> <li>Data quality - is it at an appropriate level for the consuming applications?</li> <li>How will you keep related data (in related systems) in sync?</li> <li>The case of use: Transactional (OLTP) vs Analytical (OLAP).</li> </ul> <p>Things to consider when integrating with another system:</p> <ul> <li>What's the case of use: Transactional (<a href="http://en.wikipedia.org/wiki/OLTP" rel="nofollow">OLTP</a>) or Analytical (<a href="http://en.wikipedia.org/wiki/OLAP" rel="nofollow">OLAP</a>)?</li> <li>How many requests are we talking about and when do they occur, etc?</li> <li>How complex are the queries likely to be that service these requests?</li> <li><strong>Exactly</strong> how up-to-date does the data need to be?</li> </ul> <p>If your data exists in a single DB that's serving a transactional system then any queries made against it can affect performance of that system; if you have many systems hitting that database then you've got a crucial piece of data infrastructure to protect - it goes down so does everything else.</p> <p>Databases written to support transactional systems often aren't suited to analytical ones. A common approach where I work is to ETL out copies of the data into a repository designed for the purpose in mind. The other advantage is that (as it's in a separate system) you won't get performance issues on either side.</p> <p>Assuming you do ETL out the data to a new repository which source is the authoritative one? It might be that several applications ETL data into a central <a href="http://en.wikipedia.org/wiki/Data_mart" rel="nofollow">Data Mart</a> or Data Warehouse where that data is integrated - and that becomes the "source of truth".</p> <p>In closing - working with data (both logically and physically) is a lot like Object Orientated programming / design: a lot of the underlying principles (like <a href="http://en.wikipedia.org/wiki/Solid_%28object-oriented_design%29" rel="nofollow">SOLID</a>) readily apply.</p>
 

Querying!

 
Guidance

SQuiL has stopped working due to an internal error.

If you are curious you may find further information in the browser console, which is accessible through the devtools (F12).

Reload