Table 4.
Summary of memory management studies.
| Studies | Techniques | Description | Strengths | Weaknesses |
|---|---|---|---|---|
| Optimal IP routing tables [148] | Internet backbone routing tables | Proposes the ESPRESSO heuristic to minimize the logic to compress the prefix-based match field | Proposed mechanism is helpful for effective utilization of TCAMs | Implemented and tested in traditional IP networking only |
| Effective switch memory management [151] | Floodlight, Open Daylight, OVS Switch | Based on intelligent flow management strategy to combine adaptive idle timeout values for flow rules and proactive eviction mechanism for TCAM | Beneficial for effective TCAMs utilization | Initial idle timeout, max idle timeout, and rate of timeout increase KPIs and are not considered to gauge efficiency |
| OpenFlow timeouts demystified [152] | OpenFlow 1.2, CAIDA/32 Dataset, UNIV dataset | Provides hybrid flow table management that combines timeout value with control plane eviction messages | Provides analysis of idle timeout by considering miss rate and flow table size | Dynamic setting of timeout values based on network conditions is missing |
| OPTree [153] | C++, binary search | Addresses the problem of flow rule placement in firewalls based on ACLs and reduces redundancies | Reduces the number of flow rules during flow rule placement | Lacks a consideration of the network topology change |
| Flowstat [154] | POX, Mininet | Computation of flow rules for the identified optimal paths and flow rule redistribution | Avoids congestion on network switches | Limited link failure and fault tolerance capability |
| Lossy compression of packet classifiers [157] | Gigabit ethernet Cisco 6500 switch, WireShark | Offers packet classification approach to find a classifier of optimal size to categorize the network traffic | Classify network traffic for effective TCAM usage of switches | Lacks compressing flow rules to classify a high portion of the traffic |
| Compressing forwarding tables for data center scalability [159] | TCAM, switches | Each forwarding table column is encoded separately via a dedicated variable-length binary prefix encoding | Offers a useful approach to compress forwarding tables, which is quite helpful in data center virtualization | Can be extended to investigate how other memories (CAM, TCAM) can be utilized to compress forwarding tables |