Tuning an enterprise application network becomes increasingly complex and difficult as the enterprise grows. Typically, the IT staff has increasing numbers of users, applications and locations to support along with more automated devices such as ATMs and intelligent devices. Formerly, administrators were monitoring a handful of servers, but increased usage of collaboration software and smartphone and tablet applications are exploding the network load and number of servers. Administrators are struggling to cope with the added complexity in a nonstop network that runs 365 days a year.
Get Important Answers Without Breaking Your Budget
Most IT shops face a painful dilemma: with so few good analysis resources available, dozens or hundreds of servers and hundreds or thousands of transaction types, how can network bottlenecks and problematic transactions be identified and analyzed? How can IT budget properly and assure Management that new expenditures are really necessary?
Identify Network Bottlenecks and Trouble Spots – Quickly!
Omicron’s creative use of OSIsoft PI helps enterprises zero in on application network bottlenecks and trouble spots cost-effectively and with no disruption to users or IT staff. Intermittent application network bottlenecks are nearly impossible to catch without expensive monitoring tools, but Omicron can use PI to capture and display them graphically. OSIsoft PI has an unmatched ability to continuously monitor hundreds, thousands or even tens of thousands of network, server and network data series with extraordinary data fidelity. And Omicron is masterful at using PI to give you the application network analysis answers you need.
Fast and Simple Procedure to Identify Network Bottlenecks
- Omicron meets with you to clarify your priorities and identify the network assets to be monitored.
- Omicron installs a laptop with PI on your network, configures PI and captures data for 8-10 days.
- Omicron retrieves the laptop and documents key findings over the course of one to two weeks.
- Omicron meets with you and reviews the key findings report.
For a business with 20-30 servers and 1,000 to 1,500 data series to be monitored, the typical total elapsed time is around 4 weeks. Omicron's total consultant effort is typically 4 to 5 days. Your total effort: a meeting at the beginning, a meeting at the end, and acting on the discoveries.
Omicron and PI Create Strong Benefits for Your Data Center
Performance, reliability and minimal cost: that is what Management and the user community expect from your data center operations. Omicron's application network analysis services with OSIsoft PI can help you deliver on all three.
For many data centers, the primary measure of success is the performance of the applications it supports. Yet many factors affect application performance. Sometimes, performance is largely determined by transaction rates. Other times, problems arise from configuration or hardware limitations of the database, application or webserver server. PI is able to track the response times of specific application transaction types and record key server performance statistics.
Omicron uses OSIsoft PI's strong visualization features to highlight performance issues in ways that Management can understand (box: Clearly Visualize Performance Issues).
User communities come to rely upon IT's ability to offer a stable and consistent application experience. Average response times are often built into expected productivity standards, staffing plans and time allotted to running reports for Management. Tabular data developed by Omicron using OSIsoft PI can highlight just how uneven the experiences of your users are. Uncovering such anomalies can be nearly impossible without access to a powerful real-time recording capabilities of PI.
Seeing the extreme differences between the norm (average) and the maximum time consumed by a transaction can help focus your efforts and help Management understand data center needs (box: Summarize Data in Tables).
Unless you are armed with knowledge of what's happening in your data center and network, it's tough to convince Management that you are maintaining the optimal balance between costs, performance and reliability. Justifying the purchase of another server, more memory or other hardware should rely on hard facts, not just your sales skills.
Beyond the purchase price, each additional server adds ongoing electricity costs, software support fees and maintenance efforts. Often, the costs are not necessary. A study that McKinsey & Company performed for the NY Times showed large data centers "were using only 6 to 12 percent of the electricity powering their servers to perform computations. The rest was essentially used to to keep servers idling and ready in case of a surge in activity" (Power, Pollution and the Internet, Sept. 22, 2012). Performing a study with PI can arm you with key information you need to prove you are a good steward of your corporation's funds.
Best Times for Doing a Network Analysis Study:
- Budget Preparation: Do the study early to learn how close your data center is to running out of server capacity as you add new apps and experience increased mobile usage next year.
- Merging operations: Arm yourself with dependable knowledge about the real current state of the application network you are going to merge with your own.
- Performance starts deteriorating and user complaints start increasing: Find out quickly where both constant and intermittent problems are emerging. Get ahead of it before Management starts asking about your operations.
- Technological upgrades: Make the future a little more certain by establishing a baseline before you start your upgrade, measuring the performance of your pilot and verifying the success of your rollout.
Clients call on Omicron to innovate, reduce complexity, slash maintenance and improve reliability. Call us at 1.609.678.0110 to learn how our expertise can simplify your data center tuning and budgeting efforts.