Physical - Transport - Application
|
Which layer is causing your network issues?
We understand those intermittent low downs and
network outages can cause severe impact to your
production and credibility. We work with your technical team or perform onsite analysis
to streamline your recovery process. Seeing what information is being
transferred behind the scenes can provide valuable information to underlying problems. By mapping out
your data flow we can quickly deduce the source of the issue and deliver evidence to your
vendors.
|
LONG TERM TRENDING
Those "hard to find" random problems that seem to be impossible to track down can be displayed
by our LTT product. We will tap into your network and monitor it for weeks or months. Once
the problems occurs we will track it down.
|
APPLICATION DEVELOPMENT STAGING
Keep track of where your data is going and how it is
getting there. We can work with your development team to make sure that your applications are
optimized for network communications.
|
SECURITY VERIFICATION
Are your passwords being sent in plain
text? Are your certificates working? Want to know if your
transmissions meet PCI requirements?
We can look at your data to see if it is being properly encrypted before being sent out of your
network.
|
SAMPLE ISSUES
|
Problem:
New client/server application was receiving poor response time when being leveraged in remote
sites. The number of remote clients supported did not justify upgrading to the new version of
software, thus limiting the feature sets provided by the new version and making new sales channels
difficult for staff.
Analysis Revealed:
When developing the new GUI version of the application, developers used the existing telnet protocol
communication code. This caused a large amount of excessive overhead in the transport layer of data to and from
the server. Development also created 4 separate TCP sessions per single login causing response issues when
working with 600 or more users during morning login timeframes.
Recommendation:
Change code to stop setting the "push bit" and enable the Nagle Algorithm. This would allow data to fill
the TCP packets before sending data across the WAN to optimize transmission.
Change code to leverage single server TCP session
during login.
Result:
200% gain in performance
150% gain in number of remote clients supported
|
|
Problem:
Web application using ODBC connection to a SQL Server randomly locked up during database queries for some sites while
working fine for other sites.
Analysis Revealed:
When lockups occurred, the client would send a request to the SQL server with no response
from the SQL server. Load-balancing was setup on the SQL server and not properly configured causing packets to be
sent out a different interface than the client was expecting.
Recommendation:
Reconfigure the load balancing on the SQL server. Setup Ether-channel on the switches to
accommodate MAC address resolution when clients ARP for the server information.
Result:
Information was populated to web application.
|
|
Problem:
Customer's client published a number Citrix applications to outsource work and maintain control of
staffing requirements. Citrix applications failed during peak hours when shifts over lapped causing
large production impact.
Analysis Revealed:
When high traffic was being processed by the WAN link between the customer and the client, packets were
being dropped by the customer. After review of the customer architecture it was determined that the Committed
Information Rate (CIR) from the carrier was not properly set by the customer to sustain the
additional business
from the client. Once the CIR was met, the carrier would drop the data packets causing non-response from the
client's Citrix applications.
Recommendation:
Increase the CIR with the carrier, and add a short burst time frame to account for abnormal high bandwidth
Result:
Customer was able to keep and grow additional business with client.
|
