latency, as well as a technology that scales both horizontally and vertically with high availability. This may sound like a tall order, but it’s crucial to ensuring effective data retention. Additionally, the solution must support network probes with application layer information likely to be required in future regulations, such as email not provided by CSPs, Internet telephony and other P2P services. The final and must crucial point may seem obvious, but it’s the key to compliance: the solution must efficiently store relevant retained data.

Once a data retention strategy is set, service providers must then manage this initiative cost effectively, minimize its impact on day-to-day network operations, manage and authorize warrants in a timely manner, and incorporate secure handoff of the information to law enforcement agencies.

Piece of cake, right?

Not when you consider the multitude of transactions going over a network each day. It’s almost unfathomable to think of tracking and storing each one in such a way that enables almost immediate access when a third party demands it. For starters, it’s very difficult to compile user transaction data for all activities and services. Secondly, IP traffic generates at least ten times more records than voice traffic.

So why does IP generate so many more records? While one phone call typically produces one call detail record, a single IP-based session can produce tens or hundreds of records based on the user’s plan, the time of day, and the type of data transferred. These records don’t come packaged in a convenient bundle; rather, they can arrive out of sequence and are regularly incomplete. Additionally, the number of potential identifiers for each device may be different, adding a new layer of complexity to the process.