The consulting company K-Algorithms specializes in the area of
architecture, design, development and post-implementation improvements
of complex database driven applications.
Every manager of an application development or production support team
is familiar with situations where the customer complains about their
application running slowly. When hiring new specialists for the team, questions on how to address such situations play the most
important role. And yet, diagnostics of application performance problems and/or estimation of application scalability present
a major problem which does not have a simple solution.
Performance problems can be created by one program performing an
important business function sucking all the juice from the available
hardware. Such problems can't necessarily be addressed by simply
purchasing more hardware. In particular cases such purchases would not
improve anything or would be ridiculously expensive.
In all likelihood, such programs are not using the most efficient
algorithms to get the data. In some situations it's acceptable to use a
brute force solution, in some situations it's not.
More severe problems happen when entire applications start experiencing
performance problems due to growth of data and transactional volume and
can't provide desirable levels of performance in terms of response time
and/or throughput. Again, this may be addressed by adding raw power,
using clustering and other hardware-based solutions or it may need few
touches in the data model and/or access patterns in order to break the
bottleneck in the least intrusive way.
Implementation of hardware based solutions needs to take into
consideration the underlying algorithms and application logic in order
to make best use of material investment and to get the biggest bang for
the buck. This happens for a very simple reason - even though vendors
claim that their solution is universal and provides a certain
abstraction layer, each solution has internal logic which gives the best
results only when such a solution is used properly. Specifically, it
applies to Oracle RAC, partitioning and other quite expensive solutions
which buy you nothing but administrative overhead when used incorrectly.
K-Algorithms has consultants with experience in each and every type of
database driven application you can find on the market. We have a solid
background in computer science and mathematical theory to provide
efficient solutions to your problems and extensive industrial experience
to make those solutions the most practical, least intrusive and most
profitable compared to the alternatives.
We can provide help with Oracle, DB2, MSSQL, MySQL, Informix based
applications based on any Unix/Linux or Windows platforms, different
clustering solutions, Web applications, older client-server systems,
Data Warehouses with slow ETL, cube build processes and the like.
The K(ey) is in the Algorithms!
|