MDL-based clustering for application dependency mapping

• US 10,326,672 B2
• Filed: 05/03/2016
• Issued: 06/18/2019
• Est. Priority Date: 06/05/2015
• Status: Active Grant

• Current Assignee: Cisco Technology, Inc. (Cisco Systems, Inc.)
• Original Assignee: Cisco Technology, Inc. (Cisco Systems, Inc.)
• Inventors: Scheib, Ellen Christine, Parandehgheibi, Ali, Madani, Omid, Jeyakumar, Vimalkumar, Yadav, Navindra, Alizadeh Attar, Mohammadreza
• Primary Examiner(s): Khan, Aftab N.
• Application Number:
  US15/145,666
• Publication Number:
  US 20160359697A1
• Time in Patent Office: 1,141 Days
• CPC Class Codes: G06F 16/122 using management policies backup systems G06F11/1446; file migration policies for HSM systems G06F16/185
• CPC Class Codes: G06F 16/137 Hash-based content-based indexing of textual data G06F16/31
• CPC Class Codes: G06F 16/162 Delete operations erasing in storage systems G06F3/0652
• CPC Class Codes: G06F 16/17 Details of further file system functions
• CPC Class Codes: G06F 16/173 Customisation support for file systems, e.g. localisation, multi-language support, personalisation
• CPC Class Codes: G06F 16/174 Redundancy elimination performed by the file system management of the data involved in backup or backup restore using de-duplication of the data G06F11/14
• CPC Class Codes: G06F 16/1744 using compression, e.g. sparse files
• CPC Class Codes: G06F 16/1748 De-duplication implemented within the file system, e.g. based on file segments de-duplication techniques in storage systems for the management of data blocks G06F3/0641
• CPC Class Codes: G06F 16/2322 using timestamps
• CPC Class Codes: G06F 16/235 Update request formulation
• CPC Class Codes: G06F 16/2365 Ensuring data consistency and integrity
• CPC Class Codes: G06F 16/24578 using ranking
• CPC Class Codes: G06F 16/248 Presentation of query results
• CPC Class Codes: G06F 16/285 Clustering or classification
• CPC Class Codes: G06F 16/288 Entity relationship models
• CPC Class Codes: G06F 16/29 Geographical information databases
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06F 2009/4557 Distribution of virtual machine instances; Migration and load balancing
• CPC Class Codes: G06F 2009/45587 Isolation or security of virtual machine instances
• CPC Class Codes: G06F 2009/45591 Monitoring or debugging support
• CPC Class Codes: G06F 2009/45595 : Network integration; Enabling network access in virtual machine instances
• CPC Class Codes: G06F 21/53 : by executing in a restricted environment, e.g. sandbox or secure virtual machine
• CPC Class Codes: G06F 21/552 : involving long-term monitoring or reporting
• CPC Class Codes: G06F 21/556 : involving covert channels, i.e. data leakage between processes inhibiting the analysis of circuitry or operation with measures against power attack G06F21/755
• CPC Class Codes: G06F 21/566 : Dynamic detection, i.e. detection performed at run-time, e.g. emulation, suspicious activities
• CPC Class Codes: G06F 2221/033 : Test or assess software
• CPC Class Codes: G06F 2221/2101 : Auditing as a secondary aspect
• CPC Class Codes: G06F 2221/2105 : Dual mode as a secondary aspect
• CPC Class Codes: G06F 2221/2111 : Location-sensitive, e.g. geographical location, GPS
• CPC Class Codes: G06F 2221/2115 : Third party
• CPC Class Codes: G06F 2221/2145 : Inheriting rights or properties, e.g., propagation of permissions or restrictions within a hierarchy
• CPC Class Codes: G06F 3/0482 : Interaction with lists of selectable items, e.g. menus
• CPC Class Codes: G06F 3/04842 : Selection of displayed objects or displayed text elements G06F3/0482 takes precedence
• CPC Class Codes: G06F 3/04847 : Interaction techniques to control parameter settings, e.g. interaction with sliders or dials
• CPC Class Codes: G06F 9/45558 : Hypervisor-specific management and integration aspects
• CPC Class Codes: G06N 20/00 : Machine learning
• CPC Class Codes: G06N 99/00 : Subject matter not provided for in other groups of this subclass
• CPC Class Codes: G06T 11/206 : Drawing of charts or graphs
• CPC Class Codes: H04J 3/0661 : using timestamps
• CPC Class Codes: H04J 3/14 : Monitoring arrangements for SDH/SONET rings H04J3/085
• CPC Class Codes: H04L 1/242 : by comparing a transmitted test signal with a locally generated replica
• CPC Class Codes: H04L 41/046 : comprising network management agents or mobile agents therefor
• CPC Class Codes: H04L 41/0668 : by dynamic selection of recovery network elements, e.g. replacement by the most appropriate element after failure
• CPC Class Codes: H04L 41/0803 : Configuration setting
• CPC Class Codes: H04L 41/0806 : for initial configuration or provisioning, e.g. plug-and-play
• CPC Class Codes: H04L 41/0816 : the condition being an adaptation, e.g. in response to network events
• CPC Class Codes: H04L 41/0893 : Assignment of logical groups to network elements
• CPC Class Codes: H04L 41/0894 : Policy-based network configuration management
• CPC Class Codes: H04L 41/12 : Discovery or management of network topologies
• CPC Class Codes: H04L 41/16 : using machine learning or artificial intelligence
• CPC Class Codes: H04L 41/22 : comprising specially adapted graphical user interfaces [GUI]
• CPC Class Codes: H04L 41/40 : using virtualisation of network functions or resources, e.g. SDN or NFV entities
• CPC Class Codes: H04L 43/02 : Capturing of monitoring data
• CPC Class Codes: H04L 43/026 : using flow identification
• CPC Class Codes: H04L 43/04 : Processing captured monitoring data, e.g. for logfile generation
• CPC Class Codes: H04L 43/045 : for graphical visualisation of monitoring data
• CPC Class Codes: H04L 43/062 : related to network traffic
• CPC Class Codes: H04L 43/08 : Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
• CPC Class Codes: H04L 43/0805 : by checking availability
• CPC Class Codes: H04L 43/0811 : by checking connectivity
• CPC Class Codes: H04L 43/0829 : Packet loss
• CPC Class Codes: H04L 43/0841 : Round trip packet loss
• CPC Class Codes: H04L 43/0858 : One way delays
• CPC Class Codes: H04L 43/0864 : Round trip delays
• CPC Class Codes: H04L 43/0876 : Network utilisation, e.g. volume of load or congestion level
• CPC Class Codes: H04L 43/0882 : Utilisation of link capacity
• CPC Class Codes: H04L 43/0888 : Throughput
• CPC Class Codes: H04L 43/10 : Active monitoring, e.g. heartbeat, ping or trace-route
• CPC Class Codes: H04L 43/106 : using time related information in packets, e.g. by adding timestamps
• CPC Class Codes: H04L 43/12 : Network monitoring probes
• CPC Class Codes: H04L 43/16 : Threshold monitoring
• CPC Class Codes: H04L 43/20 : the monitoring system or the monitored elements being virtualised, abstracted or software-defined entities, e.g. SDN or NFV
• CPC Class Codes: H04L 45/306 : Route determination based on the nature of the carried application
• CPC Class Codes: H04L 45/38 : Flow based routing
• CPC Class Codes: H04L 45/46 : Cluster building
• CPC Class Codes: H04L 45/507 : Label distribution
• CPC Class Codes: H04L 45/66 : Layer 2 routing, e.g. in Ethernet based MAN's
• CPC Class Codes: H04L 45/74 : Address processing for routing
• CPC Class Codes: H04L 47/11 : Identifying congestion
• CPC Class Codes: H04L 47/20 : Traffic policing
• CPC Class Codes: H04L 47/2441 : relying on flow classification, e.g. using integrated services [IntServ]
• CPC Class Codes: H04L 47/2483 : involving identification of individual flows
• CPC Class Codes: H04L 47/28 : in relation to timing considerations
• CPC Class Codes: H04L 47/31 : by tagging of packets, e.g. using discard eligibility [DE] bits
• CPC Class Codes: H04L 47/32 : by discarding or delaying data units, e.g. packets or frames
• CPC Class Codes: H04L 61/5007 : Internet protocol [IP] addresses
• CPC Class Codes: H04L 63/0227 : Filtering policies mail message filtering H04L51/212
• CPC Class Codes: H04L 63/0263 : Rule management
• CPC Class Codes: H04L 63/06 : for supporting key management in a packet data network cryptographic mechanisms or cryptographic arrangements for key management H04L9/08
• CPC Class Codes: H04L 63/0876 : based on the identity of the terminal or configuration, e.g. MAC address, hardware or software configuration or device fingerprint
• CPC Class Codes: H04L 63/1408 : by monitoring network traffic monitoring network traffic per se H04L43/00
• CPC Class Codes: H04L 63/1416 : Event detection, e.g. attack signature detection
• CPC Class Codes: H04L 63/1425 : Traffic logging, e.g. anomaly detection
• CPC Class Codes: H04L 63/1433 : Vulnerability analysis
• CPC Class Codes: H04L 63/1441 : Countermeasures against malicious traffic countermeasures against attacks on cryptographic mechanisms H04L9/002
• CPC Class Codes: H04L 63/145 : the attack involving the propagation of malware through the network, e.g. viruses, trojans or worms
• CPC Class Codes: H04L 63/1458 : Denial of Service
• CPC Class Codes: H04L 63/1466 : Active attacks involving interception, injection, modification, spoofing of data unit addresses, e.g. hijacking, packet injection or TCP sequence number attacks
• CPC Class Codes: H04L 63/16 : Implementing security features at a particular protocol layer
• CPC Class Codes: H04L 63/20 : for managing network security; network security policies in general filtering policies H04L63/0227
• CPC Class Codes: H04L 67/01 : Protocols
• CPC Class Codes: H04L 67/10 : in which an application is distributed across nodes in the network software deployment G06F8/60; multiprogramming arrangements G06F9/46
• CPC Class Codes: H04L 67/1001 : for accessing one among a plurality of replicated servers
• CPC Class Codes: H04L 67/12 : specially adapted for proprietary or special-purpose networking environments, e.g. medical networks, sensor networks, networks in vehicles or remote metering networks
• CPC Class Codes: H04L 67/51 : Discovery or management thereof, e.g. service location protocol [SLP] or web services
• CPC Class Codes: H04L 67/535 : Tracking the activity of the user network monitoring arrangements H04L43/00; recording of computer activity G06F11/34
• CPC Class Codes: H04L 67/75 : Indicating network or usage conditions on the user display
• CPC Class Codes: H04L 69/16 : Implementation or adaptation of Internet protocol [IP], of transmission control protocol [TCP] or of user datagram protocol [UDP]
• CPC Class Codes: H04L 69/22 : Parsing or analysis of headers
• CPC Class Codes: H04L 7/10 : Arrangements for initial synchronisation
• CPC Class Codes: H04L 9/0866 : involving user or device identifiers, e.g. serial number, physical or biometrical information, DNA, hand-signature or measurable physical characteristics
• CPC Class Codes: H04L 9/3239 : involving non-keyed hash functions, e.g. modification detection codes [MDCs], MD5, SHA or RIPEMD
• CPC Class Codes: H04L 9/3242 : involving keyed hash functions, e.g. message authentication codes [MACs], CBC-MAC or HMAC
• CPC Class Codes: H04W 72/54 : based on quality criteria
• CPC Class Codes: H04W 84/18 : Self-organising networks, e.g. ad-hoc networks or sensor networks

Abstract

Application dependency mapping (ADM) can be automated in a network. The network can determine an optimum number of clusters for the network using the minimum description length principle (MDL). The network can capture network and associated data using a sensor network that provides multiple perspectives and generate a graph therefrom. The nodes of the graph can include sources, destinations, and destination ports identified in the captured data, and the edges of the graph can include observed flows from the sources to the destinations at the destination ports. Each clustering can be evaluated according to an MDL score. The optimum number of clusters for the network may correspond to the number of clusters of the clustering associated with the minimum MDL score.

Claims

1. A method comprising:

capturing network flow data using sensors executing on servers of a data center network and sensors executing on networking devices connected to the servers;

determining a graph including nodes representing the servers and observed edges and unobserved edges representing the network flow data in which the observed edges between any pair of nodes of the graph indicate there are one or more observed flows between a pair of servers represented by that pair of nodes and in which the unobserved edges between any pair of nodes of the graph indicates there are no observed flows between a pair of servers represented by that pair of nodes;

determining different clusterings of the nodes of the graph in which each clustering includes clusters of one or more nodes of the graph;

determining a minimum description length (MDL) score for each clustering in which the MDL score aggregates a description length of each cluster of the clustering and in which the description length of each cluster is computed based on a minimum value between a number of the observed edges of the graph from the cluster to each other cluster of the clustering and a number of the unobserved edges of the graph from the cluster to each other cluster of the clustering;

identifying a first clustering having a minimum value for the MDL score among the different clusterings; and

generating an application dependency map including representations of applications executing in the data center network and representations of application dependencies in which the representations of applications each correspond to one of the clusters of the first clustering and the representations of application dependencies each correspond to one of the observed edges of the first clustering.

2. The method of claim 1, wherein determining the MDL score for each clustering includes:

computing the number of the observed edges from each node of each cluster of the clustering to each node of each other cluster of the clustering;

computing the of the unobserved edges between each node of each cluster of the clustering to each node of each other cluster of the clustering;

determining a minimum value between the number of the observed edges and the number of the unobserved edges for each cluster of the clustering; and

summing the minimum value between the number of the observed edges and the number of the observed edges for each cluster of the clustering.

3. The method of claim 1, wherein the data center network includes a plurality of partitions, and:

each cluster of the different clusterings consists of one or more nodes of a single partition.

4. The method of claim 1, wherein a portion of the nodes of the graph represent destination servers and destination server ports.

5. The method of claim 1, further comprising:

determining a weight for one or more of the observed edges of the graph based on a number of the observed flows corresponding to the one or more observed edges,

wherein the MDL score for each clustering is further determined based at least in part on the weight.

6. The method of claim 1, wherein each cluster of the first clustering corresponds to a respective endpoint group.

7. The method of claim 1, wherein a one of the servers of the data center network is a virtual partition, and the sensors executing on the servers of the data center network include a first sensor executing within the virtual partition and a second sensor executing on a physical server hosting the virtual partition.

8. The method of claim 1, further comprising:

capturing data in a second domain corresponding to the network flow data,

wherein a portion of the nodes of the graph represent servers and attributes of the second domain.

9. The method of claim 8, wherein the second domain is at least one of a host domain, a process domain, a user domain, a virtualization domain, or a tenant domain.

10. The method of claim 1, further comprising:

identifying an application dependency between a first application of the application dependency map and a second application of the application dependency map; and

determining a policy allowing traffic from the first application to the second application.

11. The method of claim 10, further comprising:

enforcing the policy within a simulated environment corresponding to the data center network using at least one of historical ground truth traffic data or real time ground truth traffic data.

12. A system comprising:

one or more processors; and

memory including instructions that, upon being executed by the one or more processors, cause the system to;

capture network flow data using sensors executing on servers of a data center network and sensors executing on networking devices connected to the servers;

determine a graph including nodes representing the servers and observed edges and unobserved edges representing the network flow data in which the observed edges between any pair of nodes of the graph indicate there are one or more observed flows between a pair of servers represented by that pair of nodes and in which the unobserved edges between any pair of nodes of the graph indicates there are no observed flows between a pair of servers represented by that pair of nodes;

determine different clusterings of the nodes of the graph in which each clustering includes clusters of one or more nodes of the graph; and

determine a minimum description length (MDL) score for each clustering of the different clusterings in which the MDL score aggregates a description length of each cluster of the clustering and in which the description length of each cluster is computed based on a minimum value between a number of the observed edges of the graph from the cluster to each other cluster of the clustering and a number of the unobserved edges of the graph from the cluster to each other cluster of the clustering;

identify a first clustering having a minimum value for the MDL score among the different clusterings; and

generate an application dependency map including representations of applications executing in the data center network and representations of application dependencies in which the representations of applications each correspond to one of the clusters of the first clustering and the representations of application dependencies each correspond to one of the observed edges of the first clustering.

13. The system of claim 12, wherein the instructions upon being executed further cause the system:

compute the number of the observed edges from each node of each cluster of the clustering to each node of each other cluster of the clustering;

compute the number of the unobserved edges from each node of each cluster of the clustering to each node of each other cluster of the clustering;

determine a minimum value between the number of the observed edges and the number of the unobserved edges for each cluster of the clustering; and

sum the minimum value between the number of the observed edges and the number of the unobserved edges for each cluster of the clustering.

14. The system of claim 12, wherein the data center network includes a plurality of partitions, and

each cluster of the different clusterings consists of one or more nodes of a single partition.

15. The system of claim 14, wherein the instructions upon being executed further cause the system to:

capture data in a second domain corresponding to the network flow data,

wherein a portion of the nodes of the graph represent servers and attributes of the second domain, and

wherein the second domain is at least one of a host domain, a process domain, a user domain, a virtualization domain, or a tenant domain.

16. A non-transitory computer-readable medium having computer readable instructions that, upon being executed by one or more processors, cause the one or more processors to:

capture network flow data using sensors executing on a servers of a data center network and sensors executing on networking devices connected to the servers;

determine a graph including nodes representing the servers and observed edges and unobserved edges representing the network flow data in which the observed edges between any pair of nodes of the graph indicate there are one or more observed flows between a pair of servers represented by that pair of nodes and in which the unobserved edges between any pair of nodes of the graph indicates there are no observed flows between a pair of servers represented by that pair of nodes;

determine different clusterings of the nodes of the graph in which each clustering includes clusters of one or more nodes of the graph;

determine a minimum description length (MDL) score for each clustering of the different clusterings in which the MDL score sums a description length of each cluster of the clustering and in which the description length of each cluster is computed based on a minimum value between a number of the observed edges of the graph from the cluster to each other cluster of the clustering and a number of the unobserved edges of the graph from the cluster to each other cluster of the clustering;

identify a first clustering having a minimum value for the MDL score among the different clusterings; and

generate an application dependency map including representations of applications executing in the data center network and representations of application dependencies in which the representations of applications each correspond to one of the clusters of the first clustering and the representations of application dependencies each correspond to one of the observed edges of the first clustering.

17. The non-transitory computer-readable medium of claim 16, wherein the instructions upon being executed further cause the one or more processors to:

determine a weight for one or more observed edges of the graph based on a number of the observed flows corresponding to the one or more observed edges,

wherein the MDL score for each clustering is further determined based at least in part on the weight.

18. The non-transitory computer-readable medium of claim 16, wherein the instructions upon being executed further cause the one or more processors to:

identify an application dependency between a first application of the application dependency map and a second application of the application dependency map; and

determine a policy allowing traffic from the first application to the second application.

19. The non-transitory computer-readable medium of claim 18, wherein the instructions upon being executed further cause the one or more processors to:

enforce the policy within a simulated environment corresponding to the data center network using at least one of historical ground truth traffic data or real time ground truth traffic data.