Analyzing Dynamic Network Applications: A Best Practices Overview |
By Natt Wyatt, VP Engineering, Skytide
As never before, networks are the life-blood of today’s dynamic, interactive enterprise. From the delivery of services and products to instant communications between employees, customers and prospects, today’s business is based on increasingly complex networks to conduct real-time business. These network-driven companies are more and more outward facing and customer service oriented, relying on the network to deliver new content (IPTV, video-on-demand, advertising, social networks); communication channels (VoIP, IM, wireless communications); and services (eCommerce). They attract and retain customers by offering the most reliable services and products, and/or the “coolest,” “hippest,” and most unique products available on the market.
As a result, extreme demands are placed on networking infrastructures as the volume and complexity of the network application data explodes. Companies are grappling with ways to use analytics across this massive flow of data to understand how content and services are being delivered, accessed and consumed by their customers and prospects.
Network Applications Data Analytics: What Do Companies Need to Know?
Network-dependent companies must understand the control data associated with its service offerings take place on the network. Full analysis of network traffic can help to optimize customer service and experience, while analysis of service delivery across all channels can help to competitively differentiate services. Analysis of network control data can be used to drive improvements in three critical areas:
- Network Operations — for continued optimum service delivered, and to identify new revenue opportunities
- Billing — to maximize revenue stream
- Customer Growth — drive revenue growth
This analysis, however, is increasingly difficult to obtain. Processes are constantly in flux, requiring ever more data to support. At the same time data volumes are expanding and becoming more decentralized across geographic regions and functional work groups. And, the data formats are highly diverse and “strange,” making it more difficult to correlate and combine data to obtain a holistic view.
Current analytical solutions are inadequate to resolve with these issues for three criticl reasons:
Not Scaleable |
• High volumes of network data overwhelm storage • Cube builds or queries have poor performance |
Not Flexible |
• Changes in data format requires time-consuming changes to ETL, warehouse schema and cube definitions • New queries or analytical requirements require re-architecting data warehouse structure, very laborious and time consuming • Warehouse schema changes can require heavyweight data migration |
Not Cost Effective |
• Warehouse requires large centralized infrastructure —“Big iron” solution • Licensing models prohibit “divide and conquer” strategy • Specialized expertise needed for each of ETL, warehouse, cube modeling |
Companies need an alternative to the current technologies being used to perform analytics for its netflow data.
Six Steps to Network Application Analysis: Best Practice Requirements
An effective analytical solution for today’s network-reliant company needs to meet six specific requirements. These advanced requirements are part of the next-generation of analytical software solutions that can drive smarter, faster business processes and decisions based on timely, comprehensive information.
1. Handle Vast Sources of Data
This starts with the ability to seamlessly connect to and ingest data from various data sources – network routers and applications; web servers; CRM systems; and SFA tools, to name a few. It then must allow on-the-fly alterations as new data sources are added and/or existing ones dropped. Going one step further, the system would able to automatically identify the common data types and then apply the appropriate parser, i.e.: Apache log file, tab separated file, markup language file, etc., thus speeding analytical modeling and reducing labor costs.
2. Rapid Time to Analysis
Latency is critical in analyzing network application data given the high volumes of data generated by these always-on, always-available applications and services. With terabytes or even petabyte of data/day generated, both historical analysis and ad-hoc queries for a time-limited view of the data can run into hours, days, or even longer, rendering the results meaningless. The next-generation solution will provide a way to query and view results in near real-time for both trend analysis over time and up-to-minute views to support timely decision making.
3. Easy to Deploy, Model & Query
The solution must be user-friendly for the vast majority of corporate users—not just technology experts—so business can leverage the information to support ongoing business strategies. As much as possible, the solution should be self-guiding, with “wizards” to navigate users from deployment through to creating saved analytical models and ad-hoc queries. For information to be used to its maximum potential, it must reach the business decisions makers across the entire organization—from sales and marketing to finance, support and IT, and finally to the executive level.
4. Changes can’t Break the System
The solution must readily adapt to change without requiring time-consuming and labor-intensive alterations. Today’s business environment is extremely fluid, with shifts in the market, audience development, new products, services, etc. occurring at every turn. When new data sources are identified, different views of the data are required, new data queries developed, or new products or services added to the models, the solution must deal with these changes without the need for intensive resource allocation or time-consuming re-builds of the data models.
5. Results Presented in a Meaningful Way
Analytical results must be accessible to required personnel in formats that are usable within their specific work environment. Accounting may need the sums of all sales across regions in an Excel spreadsheet, while sales may prefer this same information presented as always-current graphical charts in a web portal. Network operations may best interact with overall net-flow stats in CrystalReports, while the CTO demands access to the information as a daily roll-up report in a proprietary format. Marketing could view web traffic segmented across advertising campaigns in both up-to-date web portal views and monthly Excel spreadsheets that also combine lead qualification status. Whatever need, the solution should be able to interact with the most popular off-the-shelf presentation formats as well as customized reporting packages, with timely, always-current results.
6. Cost Effective:
Traditional analytical solutions that sit atop a large database are very expensive in terms of hardware storage and database licenses, as well as personnel costs to maintain the complex system. A true, effective analytics solution for high volumes of diverse data must reduce these costs so the system is affordable for wider use across companies of all sizes and industries.
Click here to find out how Skytide meets these challenges....
Conclusion
Business that depend on the network as a primary channel for delivering their products and services, and those that rely on the network to reach potential customers or drive internal processes, have a wealth of information hidden in the volumes of data these networks generate. Unlocking the true value of this data will be critical in sustaining revenue over time.
Find out more!
Contact Skytide today to see how Skytide can help you unlock the value hidden in your volumes of network data. Email us at info@skytide.com or call 650-292-1900.