How iWatch Works: the self-maintaining database

How can iWatch keep up?

It would be reasonable to ask how iWatch manages to keep track of such a volume of performance information, especially in busy environments where SQL transaction rates can reach into the thousands per second.

The sheer volume of data could create a high-order database administration challenge, or require costly hardware, or both. And even then, attempting to manage this data with a traditional relational database might prove unwieldy for time-related performance inquiries.

iWatch solves these challenges through the use of an industry-unique, self-maintaining, time-series database, designed from the ground up to handle large volumes of performance data. The result is a database that scales easily to handle many of thousands of SQL transactions per second, while delivering lightening-fast searches – all without the need for purging or other maintenance activities.

Using a “proprietary yet open” design philosophy, the iWatch database performs efficiently and reliably, while keeping the data open and accessible for interchange with third-party applications and industry-standard reporting tools.

Some key elements of the iWatch database architecture include:

  • Rapid Inserts. This assures that the iWatch database will keep pace with high-transaction production environments such as trading systems, even at levels exceeding 10,000 transactions per second.
  • Time-series structure. This allows the iWatch database to deliver lighting-fast responses to time related performance inquiries, a significant advantage for post event analysis (e.g. “What went wrong yesterday and 3:52 PM?”).
  • Efficient, lightweight design. The database has been designed to maintain the smallest possible footprint, even when handling query activity or significant I/O operations. This helps keep CPU overhead to 1 or 2%, allowing customers the flexibility to deploy iWatch where it’s most operationally convenient, either on a separate platform, or directly on the database host.
  • Robust filtering The iWatch database supports Perl-Compatible Regular Expressions (PCRE), allowing users to contract complex filters, either manually or aided by our intuitive GUI.
  • Flexibile retention & archiving. iWatch lets you decide just how much historical data you need to preserve. And with the optional iArchive module, you can even compress your iWatch data, and transfer it to automatically central repository – extending retention periods by 20 times or more, while reducing storage costs.
  • Open accessibility In addition to the iWatch client and web console, iWatch supports a rich array of interchange facilities to allow access from industry-standard reporting tools and third-party or custom applications. OLE DB drivers are available, along with an API, a command line export facility, and even a Data Mining Utility that can automatically extract iWatch summaries and insert the data into a traditional relational database. And any iWatch query can easily be exported to a spreadsheet.
  • And... it's included. Which means there's no hidden cost to license a repository database, and nothing else to install or configure. Install an iWatch server, plus at least one iWatch client, and you're ready to start monitoring.

Next: Querying the iWatch Data »

Previous: Capture, Parsing & Analysis «

Tell me more: request a demo or download now. »