Offline, intelligent load balancing of SCTP traffic

Abstract

Techniques for enabling offline, intelligent load balancing of Stream Control Transmission Protocol (SCTP) traffic are provided. According to one embodiment, a load balancer can receive one or more SCTP packets that have been replicated from a network being monitored. The load balancer can further recover an SCTP message from the one or more SCTP packets and can map the SCTP message to an egress port based on one or more parameters decoded from the SCTP message and one or more rules. The load balancer can then transmit the SCTP message out of the egress port towards an analytic probe or tool for analysis.

Description

CROSS-REFERENCES TO RELATED APPLICATIONS

The present application claims the benefit and priority of India Provisional Application No. 201641010295, filed Mar. 24, 2016, entitled “SYSTEM AND METHOD FOR OFFLINE LOAD BALANCING OF SCTP PROTOCOL TRAFFIC” and U.S. Pat. No. 10,999,200, filed Oct. 27, 2016, entitled “OFFLINE, INTELLIGENT LOAD BALANCING OF SCTP TRAFFIC.” The entire contents of these applications are incorporated herein by reference in their entirety for all purposes.

BACKGROUND

As known in the field of computer networking, a visibility network (sometimes referred to as a visibility fabric) is a type of network that facilitates the monitoring and analysis of traffic flowing through another network (referred to herein as a “core” network). The purposes of deploying a visibility network are varied and can include management/optimization of the core network, security monitoring of the core network, business intelligence/reporting, compliance validation, and so on.

FIG. 1 depicts an example visibility network 100 according to an embodiment. As shown, visibility network 100 includes a number of taps 102 that are deployed within a core network 104. Taps 102 are configured to replicate traffic that is exchanged between network elements in core network 104 and forward the replicated traffic to a packet broker 106 (note that, in addition to or in lieu of taps 102, one or more routers or switches in core network 104 can be tasked to replicate and forward traffic to packet broker 106 using their respective SPAN or mirror functions). Packet broker 106 can perform various packet processing functions on the traffic received from taps 102, such as removing protocol headers, filtering/classifying/correlating packets based on configured rules, and so on. Packet broker 106 can then transmit the processed traffic to one or more analytic probes/tools 108, which can carry out various types of calculations and analyses on the traffic in accordance with the business goals/purposes of visibility network 100 (e.g., calculation of key performance indicators (KPIs), detection of security threats/attacks in core network 104, generation of reports, etc.).

In cases where a single probe/tool 108 does not have sufficient capacity (e.g., compute capacity, memory capacity, storage capacity, etc) to analyze the entirety of the traffic volume replicated from core network 104, packet broker 106 can implement functionality to distribute the replicated traffic across a number of probes/tools in a load balanced manner. In this way, each individual probe/tool 108 can be tasked to handle a subset (rather than the entirety) of the replicated traffic. Existing packet brokers typically implement this load balancing functionality by calculating a hash value for each replicated packet based on a 5-tuple of packet header fields comprising <source IP address, source port, destination IP address, destination port, protocol identifier and then forwarding the packet to the probe/tool associated with the calculated hash value.

Unfortunately, while load balancing based on the foregoing 5-tuple works well for transport protocols such as TCP or UDP where traffic is always transmitted along a single path between endpoints (i.e., between a single source IP and single destination IP), it is less suitable for transport protocols such as SCTP (Stream Control Transmission Protocol) where traffic can be transmitted along one of multiple paths between endpoints (known as multi-homing). This is because multi-homing protocols support automatic failover of traffic from one path to another in response to a failure, which in the case of 5-tuple based load balancing will cause the packets for a given communication session to be hashed, and thus forwarded, to a different probe/tool after the failover than before the failover. This switch in the destination probe/tool is undesirable since all of the traffic for a single communication session (e.g., in the case of mobile network, a single mobile user session) should ideally go to the same probe/tool in order to facilitate state-based analyses.

Further, even in non-multi-homing deployments, hashing based on the 5-tuple of <source IP address, source port, destination IP address, destination port, protocol identifier> necessarily causes a given probe/tool to receive all of the traffic between the two endpoints identified in the tuple. If the volume of traffic between those two endpoints is particularly large, the probe/tool may become overloaded. Accordingly, it would be desirable to have a mechanism for performing load balancing within a visibility network that is more intelligent than simple 5-tuple hashing.

SUMMARY

Techniques for enabling offline, intelligent load balancing of Stream Control Transmission Protocol (SCTP) traffic are provided. According to one embodiment, a load balancer can receive one or more SCTP packets that have been replicated from a network being monitored. The load balancer can further recover an SCTP message from tire one or more SCTP packets and can map the SCTP message to an egress port based on one or more parameters decoded from the SCTP message and one or more rules. The load balancer can then transmit the SCTP message out of the egress port towards an analytic probe or tool for analysis.

The following detailed description and accompanying drawings provide a better understanding of the nature and advantages of particular embodiments.

BRIEF DESCRIPTION OF DRAWINGS

FIG. 1 depicts an example visibility network.

FIG. 2 depicts a visibility network comprising an SCTP load balancer according to an embodiment.

FIG. 3 depicts a diagram of two network elements that are connected via SCTP using multi-homing according to an embodiment.

FIG. 4 depicts the structure of an SCTP packet according to an embodiment.

FIGS. 5 and 6 depict a flowchart and a packet flow for performing packet modifying SCTP load balancing according to an embodiment.

FIGS. 7 and 8 depict a flowchart and a packet flow for performing packet preserving SCTP load balancing according to an embodiment.

FIG. 9 depicts an example network switch/router according to an embodiment.

FIG. 10 depicts an example computer system according to an embodiment.

DETAILED DESCRIPTION

In the following description, for purposes of explanation, numerous examples and details are set forth in order to provide an understanding of various embodiments. It will be evident, however, to one skilled in the art that certain embodiments can be practiced without some of these details, or can be practiced with modifications or equivalents thereof.

1. Overview

Embodiments of the present disclosure provide techniques for performing offline, intelligent load balancing of traffic that is transmitted via a multi-homing transport protocol such as SCTP. The qualifier “offline” means that these load balancing techniques can be performed on replicated (rather than live) traffic, such as via a packet broker in a visibility network. The qualifier “intelligent” means that these techniques can perform load balancing in a more configurable and granular fashion than 5-tuple hashing (for example, at the granularity of SCTP messages), and thus can overcome the problems that arise when using 5-tuple based load balancing in. e.g., multi-homing deployments.

These and other aspects of the present disclosure are described in further detail in the sections that follow.

2. Visibility Network

FIG. 2 depicts an example visibility network 200 that may be used to implement the intelligent load balancing techniques of the present disclosure according to an embodiment. As shown, visibility network 200 includes a number of taps 202 that are deployed in a core network 204 and are configured to replicate traffic exchanged in network 204 to a packet broker 206. In FIG. 2, core network 204 is a mobile LTE network that comprises network elements specific to this type of network, such us an eNodeB 210, a mobility management entity (MME) 212, a serving gateway (SGW) 214, and a packet data network gateway (PGW) 216 which connects to an external packet data network such as the Internet. Further, in this particular example, laps 202 are configured to replicate and forward SCTP traffic that is exchanged on certain interfaces (e.g., SIMME, SGs, S6a, Gx, and Gy) of core network 204. However, it should be appreciated that core network 204 can be any other type of computer network known in the art, such as a mobile 3G network, a landline local area network (LAN) or wide area network (WAN), etc.

Upon receiving the replicated traffic via laps 202, packet broker 206 can perform various types of packet processing functions on the traffic (as configured/assigned by an operator of visibility network 200) and can forward the processed traffic to one or more analytic probes/tools 208 for analysis. In one embodiment, packet broker 206 can be implemented solely in hardware, such as in the form of a network switch or router that relies on ASIC or FPGA-based packet processors to execute its assigned packet processing functions based on rules that are programmed into hardware memory tables (e.g., CAM tables) resident on the packet processors and/or line cards of the device. In another embodiment, packet broker 206 can be implemented solely in software that runs on. e.g., one or more general purpose physical or virtual computer systems. In yet another embodiment, pocket broker 206 can be implemented using a combination of hardware and software, such as a combination of a hardware-based basic packet broker and a software-based “session director” cluster as described in co-owned U.S. patent application Ser. No. 13/205,889, entitled “Software-based Packet Broker.” the entire contents of which are incorporated herein by reference in its entirety for all purposes.

As noted in the Background section, in cases where the replicated traffic from core network 204 needs to be load balanced across multiple probes/tools 208, conventional packet brokers typically calculate a hash value for each replicated packet based on a 5-tuple of packet header fields comprising <source IP address, source port, destination IP address, destination port, protocol identifier and forward the packet to a probe/tool mapped to the calculated hash value. However, this approach is problematic for traffic that is transported over a multi-homing transport protocol such as SCTP, since the packets for a given communication session may be inadvertently re-routed to a different probe/tool after a network failure (due to the session traffic being failed over to an alternate path between the session endpoints). Further, since 5-tuple based load balancing sends all of the traffic between two endpoint IP addresses to the same designated probe/tool, if the volume of traffic between those IP addresses becomes abnormally high, the designated probe/tool can become overloaded.

To address these and other similar issues, packet broker 206 of FIG. 2 implements a novel SCTP load balancer 218. Depending on the configuration of packet broker 206, SCTP load balancer 218 can be implemented in software, hardware, or a combination thereof. Generally speaking, SCTP load balancer 218 can receive the SCTP traffic that is replicated from core network 204 (after it has been processed via the assigned functions of packet broker 206) and can distribute that traffic across probes/tools 208 in a manner that (1) is more granular/configurable that 5-tuple hashing, and (2) ensures all of the traffic for a single communication session is sent to the same probe/tool, even if SCTP causes an automatic failover from one multi-homing path to another. In these ways, SCTP load balancer 218 can eliminate or minimize the problems associated with simple 5-tuple based load balancing.

For example, in one set of embodiments, SCTP load balancer 218 can recover SCTP messages that are embedded in the SCTP packets replicated from core network 204 and can map the individual messages to particular egress ports (and thus, probes/tools) in a load balanced manner based on user-defined rules/criteria. SCTP load balancer 218 can then transmit the SCTP messages out of the mapped egress ports in the form of newly constructed SCTP packets. This approach is referred to herein as the “pocket modifying” approach and is detailed in section (4) below.

In an alternative set of embodiments, SCTP load balancer 218 can recover SCTP messages that are embedded in the replicated SCTP packets and can map the messages to particular egress ports in a load balanced manner based on user-defined rules/criteria as noted above; however, instead of transmitting the SCTP messages in the form of newly constructed SCTP packets, SCTP load balancer 218 can transmit the messages by forwarding intact copies of the original SCTP packets (i.e., the packets received at packet broker 206 via. e.g., taps 202). In a situation where an original SCTP packet includes two messages (or portions thereof) that are mapped to two different egress ports respectively, SCTP load balancer 218 can forward a copy of that packet out of each of the two egress ports. This approach is referred to herein as the “packet preserving” approach and is detailed in section (5) below.

It should be appreciated that FIG. 2 is illustrative and not intended to limit embodiments of the present disclosure. For example, the various entities shown in FIG. 2 may be arranged according to different configurations and/or include subcomponents or functions that are not specifically described. One of ordinary skill in the art will recognize other variations, modifications, and alternatives.

3. Stream Control Transmission Protocol (SCTP)

To provide further context for the load balancing techniques described herein, the following sub-sections present a brief discussion of SCTP and its properties.

3.1 Protocol Overview and Multi-Homing

SCTP is a transport layer (i.e., OSI Layer 4) protocol that is commonly used in mobile networks such as LTE network 204 shown in FIG. 2 for signaling messages between the network's various elements (e.g., eNodeB, MME, MSC, etc.). As mentioned previously, SCTP supports multi-homing, which allows multiple, redundant paths to be denned between the endpoints of an SCTP connection (known as an SCTP “association”). In this way, SCTP can provide a level of resiliency and reliability that is not possible using alternative transport protocols such TCP or UDP.

For example, FIG. 3 depicts an MME server and a HSS server that are connected via an SCTP association comprising two paths (a first through network #1 and a second through network #2). The first path is defined between a NIC2 on the MME server having an IP address A.B.C.h and a NIC1 on the HSS server having an IP address A.B.C.j. Further, the second path is defined between a NIC1 on the MME server having an IP addresses X.Y.Z.n and a NIC2 on the HSS server having an IP address X.Y.Z.m. In a situation where a failure (e.g., port or link failure) occurs on cither path, the protocol can detect the failure can automatically redirect traffic to the other path, thereby ensuring that traffic continues to (low between the servers.

3.2 Message-Oriented Multi-Streaming

SCTP is a message-oriented protocol, which means that it transmits a sequence of messages (rather than an unbroken sequence of bytes) between endpoints. Each message is composed of a sequence of smaller units known as chunks.

SCTP is also a multi-streaming protocol, which means that it can transmit, within a single SCTP association, several independent streams of messages/chunks in parallel. Error handling is implemented on a per-stream basis and thus a packet drop, CRC error, or checksum error on one stream will not affect the transfer of other streams, which eliminates unnecessary head-of-line blocking. In LTE networks, SCTP streams are commonly used to group together messages belonging to a range of mobile users (identified by, e.g., International Mobile Subscriber Identity (IMSI)), such as one stream for every X users.

To clarify the message-oriented multi-streaming nature of SCTP, FIG. 4 depicts the structure of a typical SCTP packet 400. As shown, SCTP packet 400 includes a SCTP header and a number of chunks 1-N. Each chunk includes a stream ID which uniquely identifies the stream to which the chunk belongs, a transmission sequence number (TSN) which uniquely identifies the ordering of the chunk relative to other chunks in the same and other SCTP packets transmitted via this association, and a stream sequence number (SSN) which identifies the message within the stream to which the chunk belongs. With this information, a receiver endpoint of an SCTP association can reconstruct the streams and constituent messages sent by the sender endpoint of the association.

It should be noted that SCTP supports two types of chunks-data chunks and control chunks. Data chunks carry a message payload while control chunks are used for creating/tearing down an SCTP association, transmitting acknowledgements between endpoints, und testing reachability. To preserve message boundaries, each chunk includes a “B” (begin) hit and an “E” (end) bit; these hits indicate whether the chunk is the first chunk of a message or the last chunk of a message respectively. If both bits are set. The chunk contains the entirety of a single message.

It should also be noted that a SCTP packet may contain chunks (and thus messages) belonging to different streams. There is no requirement that a given SCTP packet comprise data solely for a single stream.

4. Packet Modifying SCTP Load Balancing

With the foregoing discussion of SCTP in mind. FIGS. 5 and 6 depict a flowchart 500 and a packet flow 600 respectively that may be carried out by SCTP load balancer 218 of FIG. 2 for performing offline, intelligent SCTP load balancing in a “packet modifying” manner according to an embodiment. These figures are described in combination below. With this packet modifying approach, SCTP load balancer 218 can recover the messages contained in the SCTP packets replicated from core network 204 and effectively repackage these messages into new SCTP packets that are sent to probes/tools 208. In this way, SCTP load balancer 218 can load balance the replicated SCTP traffic on a per-message basis towards probes/tools 208.

Starting with block 502 of flowchart 500, a packet receive module 602 of SCTP load balancer 218 (shown in packet flow 600) can receive an SCTP packet that has been replicated from core network 204 and can determine the SCTP association on which the packet was sent. Packet receive module 602 can make this determination based on. e.g., the SCTP header in the packet.

Assuming that the determined SCTP association is X, packet receive module 602 can check whether a packet queue 604(X), a packet reorder queue 606(X), and a data chunk queue 608(X) exists for association X within SCTP load balancer 218 (block 504). If not, packet receive module 602 can cause a new instance of each of these queues to be created for association X (block 506).

Upon creating queues 604(X)-608(X) (or verifying that they already exist), packet receive module 602 can further check whether the TSN of the first chunk in the SCTP pocket has continuity with (i.e., directly follows from) the last TSN received/processed by SCTP load balancer 218 for association X (block 508). If not, this means that the current SCTP packet has been received “out-of-order,” and thus packet receive module 602 can place the SCTP packet in packet reorder queue 606(X) so that it may be processed at a later point in time once the intervening packets for association X have arrived and have been processed (block 510). Flowchart 500 can then return to block 502.

However, if the TSN of the first chunk of the SCTP docs have continuity with the last TSN for association X, packet receive module 602 can enqueue the SCTP packet to packet queue 604(X) (block 512), enqueue five data chunks in the SCTP packet, in TSN order, to data chunk queue 608(X) (block 514), and pass any control chunks in the SCTP packets to a control chunk processor 610 for handling (block 516). Although not explicitly shown in flowchart 500, control chunk processor 610 can handle certain types of control chunks as indicated below:

• • INIT or INIT_ACK control chunk: processor 610 extracts endpoint IP address, starling TSN in each uplink/downlink flow, max number of streams that can be supported on each uplink/downlink Dow and associates this information with association X in an association table 612
  • SACK control chunk: processor 610 lends signal to stream processor 614 to consume packets with TSNs acknowledged by SACK chunk
  • Shutdown, shutdown ACK, or shutdown complete control chunk: processor 610 removes entry for association X from association table 612
  • Abort or error control chunk: processor 610 handles errors in connection

Note that, as part of enqueuing data chunks to data chunk queue 608 at block 514, packet receive module 602 can read from association table 612 to retrieve a pointer to queue 608.

At block 518, stream processor 614 can receive a signal from control chunk processor 610 indicating how many received SCTP packets it can consume (as noted above with respect to the SACK control chunk). In response, stream processor 614 can dequeue the data chunks for those packets from data chunk queue(s) 608, de-multiplex the data chunks based on each chunk's SID, and enqueue the de-multiplexed data chunks into a number of stream queues 616 (block 520). In this way, stream processor 614 can separate out the data chunks on a per-stream basis. As part of this process, stream processor 614 can cause the data chunks that have been dequeued from data chunk queue(s) 608 to be deleted from the corresponding packet queue(s) 604.

Further, as part of adding data chunks to stream queues 616, stream processor 614 can check the message boundaries defined the data chunks. Upon encountering the presence of the “E” bit for a given sequence of data chunks indicating a complete SCTP message, stream processor 614 can trigger a message decoder module 618 (block 522).

In response to being triggered, message decoder module 618 can parse the SCTP message using a protocol parser that is appropriate for the message (e.g., Diameter, S6a, S1AP, etc.) and can extract information from the message that is relevant to the state of the communication session to which the message belongs (block 524). In the case of a mobile user session, this information can include, e.g., user (IMSI) details, user equipment (IMEI) details, and more. Message decoder module 618 can store this state information in a state/session table 620 in order to track the states of ongoing communication sessions.

Then, at block 526, message decoder module 618 can consult an egress configuration table 622 for user-defined load balancing rules/criteria that indicate how the current message should be distributed to the egress ports of packet broker 206 (and thus, to probes/tools 208) in view or the information determined/extracted at block 524. Note that these rules/criteria are completely user-configurable and can correspond to various types of load balancing algorithms such as IMSI-based round robin, message level round robin, etc. The end result of block 526 is that message decoder module 618 can determine a specific mapping between the message and one or more specific egress ports of packet broker 206.

At block 528, message decoder module 618 can pass the message and the mapped egress port(s) to an SCTP transmit module 624. SCTP transmit module 624 can maintain pre-opened SCTP associations between packet broker 206 and each probe/tool 208. Finally, at block 530, SCTP transmit module 624 can package the message into a new SCTP packet and transmit lire newly created SCTP packet with the message out of the egress port towards a probe/tool 208.

It should be appreciated that flowchart 500 and packet flow 600 of FIGS. 5 and 6 are illustrative and various modifications are possible. For example, although the preceding description suggests that each data chunk queue 608 maintains on actual copy of the data chunks for packets added to the packet queue 604, in some embodiments each data chunk queue 608 may simply comprise pointers to the packet queue (which holds the actual data chunk data). This approach can reduce the memory footprint of the solution.

Further, while message decoder module 618 can track protocol state information and use this information to direct the load balancing process as noted above, in some embodiments module 618 may not do so. Instead, message decoder module 618 may simply decode SCTP parameters (e.g., message boundaries, stream IDs, etc.) and apply these SCTP parameters for load balancing purposes. One of ordinary skill in the art will recognize other variations, modifications, and alternatives.

5. Packet Preserving SCTP Load Balancing

As noted previously, as an alternative to the packet modifying approach of section (4). SCTP load balancer can implement a “packet preserving” approach for performing intelligent load balancing of SCTP traffic. A flowchart 700 and a packet flow 800 for this packet preserving approach are presented in FIGS. 7 and 8 respectively and described below. At a high level, the packet preserving approach is similar to the packet modifying approach hut is designed to forward load balanced SCTP messages to probes/tools 208 in the form of the original SCTP packets received at packet broker 206 (rather than in the form of newly created SCTP pockets). Thus, with this approach, there is no need to open SCTP associations between packet broker 206 and probes/tools 208.

Blocks 702-726 of flowchart 700 are generally similar to blocks 502-526 of flowchart 500, with the caveat that stream processor 614 docs not delete packets from the packet queues upon enqueuing data chunks to the stream queues. Instead, stream processor 614 can insert a reference counter into each packet in packet queue 604 indicating the number of complete messages included in that packet.

At block 728, a packet transmit module 626 that is shown in FIG. 8 (rather than SCTP transmit module 624) can receive the message and mapped egress port(s) from message decoder module 618. Then, at block 730, packet transmit module 626 can retrieve the original SCTP packet(s) corresponding to the message from packet queue(s) 604. Finally, at block 732, packet transmit module 626 can forward the original SCTP packet(s) as-is out of the mapped egress port(s) towards one or more probes/tools 208. Note that if a given packet contains multiple messages that are mapped to different egress ports, packet transmit module 626 can forward the packet multiple times (one for each different egress port).

6. Example Network Device

FIG. 9 depicts an example network device (e.g., switch/router) 900 according to an embodiment. Network switch/router 900 can be used to implement packet broker 206/SCTP load balancer 218 (or a portion thereof) according to an embodiment.

As shown, network switch/router 900 includes a management module 902, a switch fabric module 904, and a number of line cards 906(1)-906(N). Management module 902 includes one or more management CPUs 908 for managing/controlling the operation of the device. Each management CPU 908 can be a general purpose processor, such as a PowerPC, Intel, AMD, or ARM-based processor, that operates under the control of software stored in an associated memory (not shown).

Switch fabric module 904 and line cards 906(1)-906(N) collectively represent the data, or forwarding, plane of network switch/router 900. Switch fabric module 904 is configured to interconnect the various other modules of network switch/router 900. Each line card 906(1)-906(N) can include one or more ingress/egress ports 910(1)-910(N) that are used by network switch/router 900 to send and receive packets. Each line card 906(1)-906(N) can also include a packet processor 912(1)-912(N). Packet processor 912(1)-912(N) is a hardware processing component (e.g., an FPGA or ASIC) that can make wire speed decisions on how to handle incoming or outgoing traffic.

It should be appreciated that network switch/router 900 is illustrative and not intended to limit embodiments of the present invention. Many other configurations having more or fewer components than switch/router 900 are possible.

7. Example Computer System

FIG. 10 depicts an example computer system 1000 according to an embodiment. Computer system 900 can be used to implement packet broker 206/SCTP load balancer 218 (or a portion thereof) according to an embodiment.

As shown in FIG. 10, computer system 1000 can include one or more general purpose processors (e.g., CPUs) 1002 that communicate with a number of peripheral devices via a bus subsystem 1004. These peripheral devices can include a storage subsystem 1006 (comprising a memory subsystem 1008 and a file storage subsystem 1010), user interface input devices 1012, user interface output devices 1014, and a network interface subsystem 1016.

Bus subsystem 1004 can provide a mechanism for letting the various components and subsystems of computer system 1000 communicate with each other as intended. Although bus subsystem 1004 is shown schematically as a single bus, alternative embodiments of the bus subsystem can utilize multiple buses.

Network interface subsystem 1016 can serve as an interface for communicating data between computer system 1000 and other computing devices or networks. Embodiments of network interface subsystem 1016 can include wired (e.g., coaxial, twisted pair, or fiber optic Ethernet) and/or wireless (e.g., Wi-Fi, cellular, Bluetooth, etc.) interfaces.

User interface input devices 1012 can include a keyboard, pointing devices (e.g., mouse, trackball, touchpad, etc.), a scanner, a barcode scanner, a touch-screen incorporated into a display, audio input devices (e.g., voice recognition systems, microphones, etc.), and other types of input devices. In general, use of the term “input device” is intended to include all possible types of devices and mechanisms for inputting information into computer system 1000.

User interface output devices 1014 can include a display subsystem, a printer, a fax machine, or non-visual displays such as audio output devices, etc. The display subsystem can be a cathode ray lube (CRT), a flat-panel device such as a liquid crystal display (LCD), or a projection device. In general, use of the term “output device” is intended to include all possible types of devices and mechanisms for outputting information from computer system 1000.

Storage subsystem 1006 can include a memory subsystem 1008 and a file/disk storage subsystem 1010. Subsystems 1008 and 1010 represent non-transitory computer-readable storage media that can store program code and/or data that provide the functionality of various embodiments described herein.

Memory subsystem 1008 can include a number of memories including a main random access memory (RAM) 1018 for storage of instructions and data during program execution and a read-only memory (ROM) 1020 in which fixed instructions are stored. File storage subsystem 1010 can provide persistent (i.e., nonvolatile) storage for program and data files and can include a magnetic or solid-state hard disk drive, an optical drive along with associated removable media (e.g., CD-ROM, DVD, Blu-Ray, etc.), a removable (lash memory-based drive or card, and/or other types of storage media known in the art.

It should be appreciated that computer system 1000 is illustrative and not intended to limit embodiments of the present disclosure. Many other configurations having more or fewer components than computer system 1000 are possible.

The above description illustrates various embodiments of the present disclosure along with examples of how aspects of the present disclosure may be implemented. The above examples and embodiments should not be deemed to be the only embodiments, and are presented to illustrate the flexibility and advantages of the present disclosure as defined by the following claims. For example, although certain embodiments have been described with respect to particular workflows and steps, it should be apparent to those skilled in the art that the scope of the present disclosure is not strictly limited to the described workflows and steps. Steps described as sequential may be executed in parallel, order of steps may be varied, and steps may be modified, combined, added, or omitted. As another example, although certain embodiments have been described using a particular combination of hardware and software, it should be recognized that other combinations of hardware and software are possible, and that specific operations described as being implemented in software can also be implemented in hardware and vice versa.

The specification and drawings are, accordingly, to be regarded in an illustrative rather than restrictive sense. Other arrangements, embodiments, implementations and equivalents will be evident to those skilled in the art and may be employed without departing from the spirit and scope of the invention as set forth in the following claims.

Claims

1. A method comprising: receiving a Stream Control Transmission Protocol (SCTP) packet that has been replicated from a network being monitored, wherein the SCTP packet comprises a plurality of data chunks including a first data chunk and another data chunk, wherein the first data chunk includes a first transmission sequence number, wherein the first data chunk carries a payload for an SCTP message, wherein the SCTP message is mapped to a first egress port, and wherein the another data chunk carries another payload for another SCTP message, wherein the another SCTP message is mapped to a second egress port;determining an SCTP association for the SCTP packet, wherein the SCTP association comprises a first endpoint and a second endpoint of an SCTP connection and wherein the first transmission sequence number identifies an ordering of the first data chunk relative to other data chunks within the SCTP association;selecting a packet queue and a data chunk queue based on the SCTP association;placing the SCTP packet into the packet queue;placing the first data chunk into the data chunk queue based on the first transmission sequence number;detecting a defined message boundary in the first data chunk of the plurality of data chunks, wherein the defined message boundary indicates a boundary of the SCTP message;triggering a message decoder module based on detecting the defined message boundary;determining, by the message decoder module, a mapping of the SCTP message to the first egress port and a second mapping of the another SCTP message to the second egress port; andtransmitting, based on the mapping, a first copy of the SCTP packet via the first egress port and, based on the second mapping, a second copy of the SCTP packet via the second egress port.

2. The method of claim 1, wherein the message decoder module is triggered based on the boundary of the SCTP message indicating that the first data chunk is a last chunk of the SCTP message.

3. The method of claim 1, further comprising: selecting a packet reorder queue based on the SCTP association;determining that the SCTP packet is received out of order based on the first transmission sequence number in an SCTP header of the SCTP packet; andplacing the SCTP packet in the packet reorder queue based on the determining.

4. The method of claim 3, wherein the SCTP association is determined based on information in the SCTP header of the SCTP packet.

5. The method of claim 1, wherein the first data chunk further includes a stream identifier and a stream sequence number, wherein the stream identifier identifies a stream to which the first data chunk belongs and wherein the stream sequence number identifies a message within the stream to which the first data chunk belongs.

6. The method of claim 1, wherein the SCTP packet further includes a control chunk, the method further comprising: passing the control chunk to a control chunk processor;receiving, from the control chunk processor, a signal for processing the SCTP packet, wherein the signal is generated based on a type of the control chunk; andprocessing the plurality of data chunks of the SCTP packet based on the signal.

7. The method of claim 1, wherein the plurality of data chunks comprises a second data chunk, wherein the second data chunk includes a second transmission sequence number and wherein the second data chunk carries a second payload for the SCTP message.

8. The method of claim 7, wherein the another data chunk includes a third transmission sequence number.

9. The method of claim 8, wherein the SCTP message is associated with a first stream and the another SCTP message is associated with a second stream.

10. The method of claim 9, further comprising: placing the first data chunk and the second data chunk into a first stream queue, wherein the first stream queue is associated with the first stream; andplacing the another data chunk into a second stream queue, wherein the second stream queue is associated with the second stream.

11. The method of claim 1 wherein the transmitting comprises: retrieving, from the packet queue, the first copy of the SCTP packet that contains the SCTP message.

12. The method of claim 1, wherein another SCTP packet has been replicated from the network being monitored, the method further comprising: receiving the another SCTP packet, wherein the SCTP packet and the another SCTP packet originate from different SCTP associations in the network being monitored.

13. The method of claim 1, wherein mapping the SCTP message comprises: extracting session information from the SCTP message; anddetecting, based on the session information, a load-balancing rule that maps the SCTP message to the first egress port.

14. The method of claim 1, wherein the defined message boundary comprises at least one of a begin bit and an end bit, wherein the begin bit indicates that the first data chunk is a first chunk of the SCTP message and the end bit indicates that the first data chunk is a last chunk of the SCTP message.

15. A non-transitory computer readable storage medium having stored thereon program code that when executed by a load balancer processor cause the load balancer processor to perform operations comprising: receiving a Stream Control Transmission Protocol (SCTP) packet that has been replicated from a network being monitored, wherein the SCTP packet comprises a plurality of data chunks including a first data chunk and another data chunk, wherein the first data chunk includes a first transmission sequence number and wherein the first data chunk carries a payload for an SCTP message, wherein the SCTP message is mapped to a first egress port, and wherein the another data chunk carries another payload for a second SCTP message, wherein the second SCTP message is mapped to a second egress port;determining an SCTP association for the SCTP packet, wherein the SCTP association comprises a first endpoint and a second endpoint of an SCTP connection and wherein the first transmission sequence number identifies an ordering of the first data chunk relative to other data chunks within the SCTP association;selecting a packet queue and a data chunk queue based on the SCTP association;placing the SCTP packet into the packet queue;placing the first data chunk into the data chunk queue based on the first transmission sequence number;detecting a defined message boundary in the first data chunk of the plurality of data chunks, wherein the defined message boundary indicates a boundary of the SCTP message;triggering a message decoder module based on detecting the defined message boundary;determining, by the message decoder module, a mapping of the SCTP message to the first egress port and a second mapping of the another SCTP message to the second egress port; andtransmitting, based on the mapping, a first copy of the SCTP packet via the first egress port and, based on the second mapping, a second copy of the SCTP packet via the second egress port.

16. The non-transitory computer readable storage medium of claim 15, wherein the SCTP packet further includes a control chunk, the operations further comprising: passing the control chunk to a control chunk processor;receiving, from the control chunk processor, a signal for processing the SCTP packet, wherein the signal is generated based on a type of the control chunk; andprocessing the plurality of data chunks of the SCTP packet based on the signal.

17. The non-transitory computer readable storage medium of claim 15, wherein the plurality of data chunks comprises a second data chunk, wherein the second data chunk includes a second transmission sequence number and wherein the second data chunk carries a second payload for the SCTP message.

18. A load balancing system comprising: a processor; anda non-transitory computer readable medium having stored thereon program code that, when executed by the processor, causes the processor to: receive a Stream Control Transmission Protocol (SCTP) packet that has been replicated from a network being monitored, wherein the SCTP packet comprises a plurality of data chunks including a first data chunk and another data chunk, wherein the first data chunk includes a first transmission sequence number and wherein the first data chunk carries a payload for an SCTP message, wherein the SCTP message is mapped to a first egress port, and wherein the another data chunk carries another payload for a second SCTP message, wherein the second SCTP message is mapped to a second egress port;determine an SCTP association for the SCTP packet, wherein the SCTP association comprises a first endpoint and a second endpoint of an SCTP connection and wherein the first transmission sequence number identifies an ordering of the first data chunk relative to other data chunks within the SCTP association;select a packet queue and a data chunk queue based on the SCTP association;place the SCTP packet into the packet queue;place the first data chunk into the data chunk queue based on the first transmission sequence number;detect a defined message boundary in the first data chunk of the plurality of data chunks, wherein the defined message boundary indicates a boundary of the SCTP message;trigger a message decoder module based on detecting the defined message boundary;determine, by the message decoder module, a mapping of the SCTP message to the first egress port and a second mapping of the another SCTP message to the second egress port; andtransmit, based on the mapping, a first copy of the SCTP packet via the first egress port and, based on the second mapping, a second copy of the SCTP packet via the second egress port.

19. The load balancing system of claim 18, wherein the SCTP packet further includes a control chunk, the processor further configured to: pass the control chunk to a control chunk processor;receive, from the control chunk processor, a signal for processing the SCTP packet, wherein the signal is generated based on a type of the control chunk; andprocess the plurality of data chunks of the SCTP packet based on the signal.

20. The load balancing system of claim 18, wherein the plurality of data chunks comprises a second data chunk, wherein the second data chunk includes a second transmission sequence number and wherein the second data chunk carries a second payload for the SCTP message.