Memorial Day weekend was relaxing. It was a nice break from constant business travel. Besides golfing, surfing, and spending quality time with the family, I also got a chance to reflect on my recent discussions with customers and peers regarding the direction of network and application performance management. One common thread emerged from these discussions: We need a top-down approach to performance management in order to achieve reduced MTTR.
Don’t get me wrong. I understand the importance and temptation of looking at the packets. However, it simply doesn’t make sense and is not scalable to begin with thousands/millions of packets at the bottom. It is like finding a needle in the haystack! A better and more efficient approach is to start with flow data on the top to view performance information such as users and applications. From this network and application flow vantage point, you can then drill down to do targeted data mining and analysis on the packet level.
With this approach, you can better address performance issues in these three areas:
- Enterprise-wide real-time visibility integrated with packet level history
- Capacity planning, trending and baselining analysis etc.
- Integration of reactive and proactive technologies to achieve more automation
I will write more entries to discuss each of these topics in more detail. In the meantime, please send me your comments and observations.
Comments