METHOD AND AN APPARATUS TO A CLOUD MANAGEMENT SYSTEM ALLOCATING DYNAMICALLY WAN OPTIMIZATION FUNCTION
RESOURCES
BACKGROUND
Field
[0001] The present application is related generally to wireless protocols, and more specifically, to configuration of the demodulation reference signal in wireless networks.
Related Art
[0002] In a related art hybrid cloud environment, which is composed of a provider's cloud environment and tenants' on-premise environment, the use of virtual machine (VM) migration technology is increasing to balance work load among multiple server hosts. For the migration, layer-2 (L2) tunnels between a source host and a destination host are used to transfer traffic for the migration as shown in FIG. 1. In a related art example, there is a storage system that provides virtual ports, and is able to transfer the virtual ports among physical ports located on multiple storage control units making up the storage system.
[0003] Related art technology also realized the virtualization of multiple physical storage systems. In the related art, storage administrators can deploy multiple physical storage systems and migrate a connection between a host and a storage system from a physical storage port to another port of a different storage system by virtualization of a set of physical ports with a virtual port. The host can use the new physical port without the change of configuration and disruption of storage access.
[0004] In an information technology (IT) environment composed of multiple sites connected with a wide-area network (WAN), the WAN may cause decreased performance of transmission control protocol (TCP) traffic across the WAN. To address the issue, WAN Optimization Devices (WODs) are used as shown in FIG. 2. The WODs are network devices that utilize several methods such as a TCP-window control, caching, de- duplication, and compression to improve the TCP performance. In a related art example, there is a system composed of multiple WODs deployed to multiple customer sites (offices) and a cloud-service provider site.
[0005] By combining the above related art examples, the TCP performance issue in hybrid cloud environments can be addressed. However, there is an issue regarding storage access performance. Because the cost of a WOD is generally higher than network switches or routers, the allocated resources of WODs are less than the resources of the switches or routers. Thus, if all traffic across the WAN is processed by WODs, actual throughput of the WAN is decreased compared to that of switches/routers. Further, important traffic can be delayed by low-priority traffic.
SUMMARY [0006] Aspects of the present application include a management computer, which may include a memory and a processor. The processor may be configured to retrieve first information associated with a network device from a plurality of network devices, each having a function for optimizing network traffic and being managed by using the memory, based on second information indicating that an object is one of copied and migrated from a source computer to a destination computer, wherein the network device of the plurality of network devices is configured to modify network traffic of data flow from the source computer to the destination computer so that third information regarding data flow is registered to the network device of the plurality of network devices. [0007] Aspects of the present application include a computer program containing instructions. The instructions may include retrieving first information associated with a network device from a plurality of network devices, each having a function for optimizing network traffic and being managed by using the memory, based on second information indicating that an object is one of copied and migrated from a source computer to a destination computer, wherein the network device of the plurality of network devices is configured to modify network traffic of data flow from the source computer to the destination computer so that third information regarding data flow is registered to the network device of the plurality of network devices. The instructions may be stored in a computer readable storage medium, or a computer readable signal medium.
[0008] Aspects of the present application may include a system involving a plurality of network devices, and a management computer containing a memory and a processor. The processor may be configured to retrieve first information associated with a network device from a plurality of network devices, each having a function for optimizing network traffic and being managed by using the memory, based on second information indicating that an object is one of copied and migrated from a source computer to a destination computer, wherein the network device of the plurality of network devices is configured to modify network traffic of data flow from the source computer to the destination computer so that third information regarding data flow is registered to the network device of the plurality of network devices.
BRIEF DESCRIPTION OF THE DRAWINGS
[0009] FIG. 1 illustrates an example of a related art hybrid cloud system that can migrate VMs among server hosts via a L2 tunnel. [0010] FIG. 2 illustrates an example of a related art network using WODs to improve the performance of network transfer across a WAN.
[0011] FIG. 3 illustrates an example of a hybrid cloud system adopting WODs to improve the storage access performance across a WAN, in accordance with an example implementation.
[0012] FIG. 4 illustrates an example of the architecture of the cloud network manager, in accordance with an example implementation.
[0013] FIG. 5 illustrates an example of the target flow information stored on the cloud network manager, in accordance with an example implementation. [0014] FIG. 6 illustrates an example of the host location stored on the cloud network manager, in accordance with an example implementation.
[0015] FIG. 7 illustrates an example of the preset distance configuration stored on the cloud network manager, in accordance with an example implementation.
[0016] FIG. 8 illustrates an example of the RTT monitoring information stored on the cloud network manager, in accordance with an example implementation.
[0017] FIG. 9 illustrates an example of the preset RTT configuration stored on the cloud network manager, in accordance with an example implementation.
[0018] FIG. 10 shows an example of the architecture of a WOD, in accordance with an example implementation. [0019] FIG. 11 shows an example of target flow information of a WOD, in accordance with an example implementation.
[0020] FIG. 12 shows an example of target flow information of a WOD, in accordance with an example implementation. [0021] FIG. 13 illustrates an example of flow chart of a WOD set-up process of the cloud network manager, in accordance with an example implementation.
[0022] FIG. 14 illustrates an example of network wherein both VM migration and storage migration are utilized, in accordance with an example implementation.
[0023] FIG. 15 illustrates an example of the target information of the cloud network manager, in accordance with an example implementation.
[0024] FIG. 16 illustrates an example of the target flow information of the WOD (wodl), in accordance with an example implementation.
[0025] FIG. 17 illustrates an example of the target flow information of the WOD (wod2), in accordance with an example implementation.
[0026] FIG. 18 illustrates an example of the volume migration information of the cloud network manager, in accordance with an example implementation.
[0027] FIG. 19 illustrates an example of the storage controller information of the cloud network manager, in accordance with an example implementation.
[0028] FIG. 20 illustrates an example of the flow chart of the cloud network manager to allocate WOD resources for VM migration, in accordance with an example
implementation.
[0029] FIG. 21 illustrates an example of a network that contains a cloud network manager which has a capability to allocate WOD resources dynamically, in accordance with an example implementation.
[0030] FIG. 22 illustrates an example of the unknown flow information composed of the flow information that is sent from the WODs and is received by the cloud network manager. [0031] FIG. 23 illustrates an example of the flow information of the cloud network manager, in accordance with an example implementation.
[0032] FIG. 24 illustrates an example of the replication information which the storage manager stores to manage the configurations of volume replication between two hosts, in accordance with an example implementation.
[0033] FIG. 25 illustrates an example of the flow chart of the set-up process of the cloud network manager, in accordance with an example implementation.
[0034] FIG. 26 illustrates an example of the iSCSI WRITE check flow, in accordance with an example implementation.
DETAILED DESCRIPTION
[0035] Some example implementations are described with reference to drawings. The example implementations that are described herein do not restrict the inventive concept, and one or more elements that are described in the example implementations may not be essential for implementing the inventive concept. Further, although certain elements may be referred to in the singular form, the elements are not intended to be limited to the singular and may also be implemented with one or more of the same element, depending on the desired implementation. The elements may be connected physically or
communicatively (e.g., wireless), depending on the desired implementation.
[0036] In the following descriptions, the process is described while a program is handled as a subject in some cases. For a program executed by a processor, the program executes the predetermined processing operations. Consequently, the program being processed can also be a processor. The processing that is disclosed while a program is handled as a subject can also be a process that is executed by a processor that executes the program or an apparatus that is provided with the processor (for example, a control device, a controller, and a storage system). Moreover, a part or a whole of a process that is executed when the processor executes a program can also be executed by a hardware circuit as substitute for or in addition to a processor.
[0037] The instructions for the program may be stored in a computer readable storage medium, which includes tangible media such as flash memory, random access memory (RAM), Hard Disk Drive (HDD) and the like. Alternatively, instructions may be stored in the form of a computer readable signal medium, which includes other media such as carrier waves.
[0038] FIG. 1 illustrates an example of a related art hybrid cloud system that can migrate VMs 102 and 109 among server hosts 101 and 108 via a L2 tunnel 114. Host 101, the onsite system, contains VM 102 and virtual switch (VSW) 103, and connects to local area network (LAN) 105, router 106 and WAN 107. Host 108, the cloud system, contains VM 109 and VSW 110, and connects to LAN 112, router 113 and WAN 107. LANs 105 and 112 connect to datastores 104 and 111, respectively.
[0039] FIG. 2 illustrates an example of a related art network using WODs to improve the performance of network transfer across a WAN. Host 201 connects to LAN 203, WOD 204, router 205 and WAN 206. Host 207 connects to LAN 209, WOD 210, router 211 and WAN 206. LANs 203 and 209 connect to datastores 202 and 208, respectively.
[0040] From the related art examples described above and as shown in FIGS. 1 and 2, the example implementations of the present application are directed to the above described problem and to improve system availability of multiple link systems that are connected to a network fabric. In the example implementations, one or more WODs are allocated to modify the network traffic of data flow (e.g., optimize or otherwise improve the traffic flow), as described below. Though the example implementations may involve optimizing the traffic flow, the present application is not limited as such, and the traffic flow may be improved or otherwise configured by the WODs depending the desired implementation.
First example implementation (VM migration based WOD resource allocation)
[0041] The first example implementation involves a hybrid cloud management system that allocates WOD resources for storage access traffic between a host and a network-attached storage (NAS) after VM migration across the WAN occurs.
[0042] FIG. 3 illustrates an example of a hybrid cloud system adopting WODs to improve the storage access performance across a WAN 308, in accordance with an example implementation. FIG. 3 involves a situation wherein a system administrator migrates a VM 302 from an onsite host 301 (e.g., a source computer) in a site to a cloud host 309 (e.g., a destination computer) in another site with a VM manager 315. Host 301, which contains VM 302 and VSW 303, is connected to LAN 305 which connects to WOD 306, router 307 and WAN 308. Host 309, which contains VM 310 and VSW 311, connects to LAN 312, WOD 313, router 314 and WAN 308. The migrated VM uses a Network Attached Storage (NAS) for its datastore. The migrated VM accesses the NAS via the network file system (NFS) 304 across the WAN 308 after its migration. The cloud network manager 316 is notified of the VM migration from the VM manager 315 by, for example, by a Simple Network Management Protocol (SNMP) trap. [0043] FIG. 4 illustrates an example of the architecture of the cloud network manager 316 connected to Management LAN 710, in accordance with an example implementation. Cloud network manager 316 stores target flow information 409, host location information 410, preset distance configuration 411, round trip time (RTT) monitoring information 412, and preset RTT configuration 413 in memory 406. Cloud network manager 316 also stores and executes a hybrid cloud control program 408 and operating system (OS) 407.
Memory 406 is connected to central processing unit (CPU) 402, input/output interface (I/O) 403, network interface controller (NIC) 404, and Storage 405. Memory 406 and storage 405 may take the form of a computer readable storage medium or can be replaced by a computer readable signal medium. Cloud network manager 316 may be implemented in the form of a management computer.
[0044] FIG. 5 illustrates an example of the target flow information 409 stored on the cloud network manager 316, in accordance with an example implementation. Target flow information 409 can be implemented as a table wherein each entry is composed of WOD ID, destination IP address, source IP address, destination TCP/UDP port number, source TCP/UDP port number, and optimization status. Each entry of this table describes a TCP/UDP flow managed by the cloud network manager 316 of the example
implementations and the corresponding WOD as the acceleration-enforcement point for the flow. For example, the first entry, indicates at a WOD associated with the onsite system (wodl) it is configured to accelerate upward traffic from the onsite system to the cloud. The second entry, indicates that a WOD associated with the cloud system (wod2) accelerates downward traffic from the cloud to the onsite system.
[0045] FIG. 6 illustrates an example of the host location information 410 stored on the cloud network manager 316, in accordance with an example implementation. The host location 410 can be implemented as a table wherein each entry is composed of host ID, location ID, and coordinates of the location. Each entry of this table describes a location of a host. For example, the first entry and the second entry of this table are equivalent to the host 301 and the host 309, respectively. Also, the third entry is equivalent to the NAS 304. [0046] FIG. 7 illustrates an example of the preset distance configuration 411 stored on the cloud network manager 316, in accordance with an example implementation. The preset distance configuration 41 1 can be implemented as a variable that stores distance value. This value represents a maximum distance between two sites which communicate without WODs. The preset distance configuration 411 can be adjusted according to a desired implementation. This distance configuration 411 and the above host location information 410 are utilized by the network manager 316 to determine whether to apply acceleration. For example, because the distance between VM 310 and NAS 304 is more than 1000km, the network manager 316 applies acceleration to the flow. [0047] FIG. 8 illustrates an example of the RTT monitoring information 412 stored on the cloud network manager 316, in accordance with an example implementation. RTT monitoring information 412 can be implemented as a table wherein each entry is composed of two location IDs and the RTT measured between two location ID sites. The network manager 316 measures periodically RTT between two sites and updates the RTT monitoring information 412.
[0048] FIG. 9 illustrates an example of the preset RTT configuration 413 stored on the cloud network manager 316, in accordance with an example implementation. The preset RTT configuration 413 can be implemented as a variable that stores a time value. This value represents a maximum RTT between two sites communicating without WODs. [0049] FIG. 10 illustrates an example of the architecture of a WOD network device 1001, in accordance with an example implementation. WOD 1001 may include a controller 1002, a backplane switch 1003, routing module 1004, and WAN optimization module 1005. The controller 1002 executes a management program 1013 and stores target flow information 1012 and has OS 1011 on memory 1009 which connects to CPU 1006, I/O 1007, NIC 1008 and Storage 1010. The routing module 1004 contains a routing engine 1014, flow configuration 1015, routing table 1016, and multiple Media Access Control Physical Layers (MAC/PHYs) 1017-1018. The management program 1013 of the controller 1002 configures the flow configuration 1015 on the routing module 1004 according to the target flow information 1012 when it is updated by the cloud network manager 316. The WAN optimization module 1005 contains a WAN optimization engine 1019 which processes packets transferred from the routing module 1004. WAN
optimization module 1005 executes TCP-window control, caching, de-duplication, or compression before it transfers the packets to another WOD. [0050] FIG. 11 illustrates an example of target flow information 1012 of a WOD 306, in accordance with an example implementation. Each entry is composed of a destination IP address, a source IP address, a destination TCP/UDP port, a source TCP/UDP port, and optimization status of the flow. The first entry represents that NFS requests from a host 309 to the NAS 304 are optimized with a WOD 313. The cloud network manager 316 creates entries of this target flow information 1012 based on the entries of its target flow information 409. In this example, the first entry of the target flow information 409 is configured to the target flow information 1012 of the WOD 306.
[0051] FIG. 12 illustrates an example of target flow information 1012 of a WOD 313, in accordance with an example implementation. The target flow information 1012 has a same structure with the above described target flow information 1012 of the WOD 306. The first entry represents that NFS responses from the NAS 304 to host 309 are optimized with a WOD 306. As well as the case of the target flow information 1012 of the WOD 306, the cloud network manager 316 creates entries of this target flow information 1012 of the WOD 313 based on the entries of its target flow information 409. In this example, the second entry of the target flow information 409 is configured to the target flow
information 1012 of the WOD 313.
[0052] FIG. 13 illustrates an example of flow chart of a WOD set-up process of the cloud network manager 316, in accordance with an example implementation. The WOD set-up process begins when the cloud network manager 316 receives a notification of VM migration from a VM manager 315. The cloud network manager retrieves the source and destination hosts of the VM migration at 1301 and retrieves the location of the migrated VM at 1302. At 1303, the cloud network manager retrieves the datastore of the migrated VM, and retrieves the location of the datastore (e.g. the onsite NFS datastore) at 1304. By accessing the VM manager 315 and retrieving host location information 410, the cloud network manager calculates the distance between the destination host and the datastore at 1305 that the migrated VM is using. The distance can be determined based on the RTT between the host and the datastore. If the calculated distance is longer than the preset distance at 1306 (Y), the cloud network manager retrieves, from a VM manager 315, a protocol that the migrated VM is using to access the datastore at 1307. Further, the cloud network manager retrieves a closer WOD located on the route at 1308 between the destination host and the datastore and then registers a flow entry at 1309 on the target flow information 409, if the calculated distance is not longer than the preset distance (N), no operation is taken at 1310. [0053] By the above described flow, especially by the flow that the cloud manager 316 analyzes the notification from the VM manager 315 immediately after the VM migration and creates the target-flow definition for WAN optimization, the example implementation can allow users to allocate efficiently limited resources of WAN optimization devices selectively for the specific use such as the remote NAS access from the migrated VM. Second example implementation (Storage volume migration based WOD resource allocation)
[0054] The second example implementation involves a hybrid cloud management system that allocates WOD resources for storage access traffic between two storage volumes after the storage systems are configured to migrate a storage volume between them.
[0055] FIG. 14 illustrates an example of a network wherein both VM migration and storage migration are utilized, in accordance with an example implementation. In this example, the storage is migrated after or simultaneously with the VM migration. For the migration, the two storage systems 1403 and 1413 use Internet Small Computer System Interfaces (iSCSI) to transfer data. Once the storage migration is completed, the host 1410 accesses the storage system 1413 instead of the storage system 1403. The host reads data from a volume 1415 via a controller 1414 if the requested data is synched between two storage systems. Until the storage migration is completed, the storage system 1413 access the storage system 1403 to obtain the data that is not synched between the two storage systems. Storage system 1403 reads a volume 1405 via a controller 1404. Host 1410, which contains VM 1411 and VSW 1412, is connected to LAN 1416, WOD 1417, router 1416, and WAN 1409. Host 1401, which contains VSW 1402, is connected to LAN 1406, WOD 1407, router 1408 and WAN 1409. Storage Manager 1417 includes volume migration information 1421 and storage controller information 1422 and is connected to LAN 1419. Cloud Network Manager 1418 is also connected to LAN 1419 which connects to router 1420 and WAN 1409.
[0056] To allocate WOD resources dynamically for the intermediate storage access between storage systems, the cloud network manager 1418 interacts with a storage manager 1417 which manages the volume migration status with volume migration information 1421 and storage controllers with storage controller information 1422. It has the same architecture with the cloud network manager 316 described in the first example implementation. Their differences are only within the hybrid cloud control program 408. The differences are described below. [0057] FIG. 15 illustrates an example of the target flow information 409 of the cloud network manager 1418, in accordance with an example implementation. In this example implementation, because the storage system 1413 on cloud accesses the data in the storage system 1403 onsite via iSCSI, target flow information 409 stores two entries that represent iSCSI response flow and iSCSI request flow. The iSCSI response flow is optimized at the WOD 1407 and the iSCSI request flow is optimized at the WOD 1417.
[0058] FIG. 16 illustrates an example of the target flow information of the WOD 1407 (wodl) 1012, in accordance with an example implementation. In this example
implementation, target flow information of wodl 1012 is a subset of the first entry of the above described target flow information 409 of the cloud network manager 1418. The first entry of this table represents a flow of iSCSI responses from the storage system 1403 to the storage system 1413.
[0059] FIG. 17 illustrates an example of the target flow information of the WOD 1417 (wod2) 1012. In this example implementation, target flow information of wod2 1012 is a subset of the second entry of the above descried target flow information 409 of the cloud network manager 1418. The first entry of this table represents a flow of iSCSI requests from the storage system 1413 to the storage system 1403.
[0060] FIG. 18 illustrates an example of the volume migration information 1421 of the cloud network manager 1418, in accordance with an example implementation. Each entry of the volume migration information 1421 represents a migration of a volume between two storage systems. Each entry includes a storage system ID, a volume ID, and controller ID for each of a destination storage system and a source storage system. The controller ID represents a storage system's controller used to handle data stream of volume migration. In this example, the first entry of the table represents storage volume migration from the storage system 1403 to the storage system 1413. More specifically, a volume 1405 on the storage system 1403 is migrated to a volume 1415 on the storage system 1413.
[0061] FIG. 19 illustrates an example of the storage controller information 1422 of the cloud network manager 1418, in accordance with an example implementation. Each entry represents a storage controller of the system. Each entry includes a storage system ID, a controller ID, protocol type, an IP address of a controller, and a World Wide Name
(WWN) of a controller. Depending on the protocol used by the controller, one of the IP address or the WWN is filled. In this example, because two storage systems use iSCSI for volume migration, iSCSI controller information is listed up in the table.
[0062] FIG. 20 illustrates an example of the flow chart of the hybrid cloud control program 408 of the cloud network manager 1418 to allocate WOD resources for VM migration, in accordance with an example implementation. The cloud network manager 1418 starts the allocation process when it receives a storage volume migration notification from the storage manager 1417 at 2001. Contents of the notification are equivalent to the changed entries in the volume migration information 1421. Then it retrieves storage controllers at 2002 and IP addresses at 2003 that are used for the storage migration. It retrieves the storage controllers from the volume migration information 1421. Also, it retrieves the IP addresses from the storage controller information 1422. The cloud network manager 1418 retrieves two WODs on a route between the controllers using network topology information 414 at 2004. The cloud network manager 1418 then creates and registers the iSCSI request flow at 2005 and the iSCSI response flow at 2006. The first entry of the target information 409 of the cloud network manager 1418 and the first entry of the target flow information 1012 of WOD 1407 (wodl) are both equivalent to the iSCSI response flow. The second entry of the target information 409 of the cloud network manager 1418 and the first entry of the target flow information 1012 of WOD 1417 (wod2) are both equivalent to the iSCSI request flow.
[0063] By the above described flow, especially by the flow that the cloud manager 1418 analyzes the notification from the storage manager 1417 immediately after the activation of volume migration and creates the target-flow definition for WAN optimization, the example implementation can allow users to allocate efficiently limited resources of WAN optimization devices selectively for the specific use such as the storage volume access between multiple storage systems that are configured to migrate a volume from one side to another side.
Third example implementation (Storage volume migration based dynamic WOD resource allocation)
[0064] The third example implementation involves a hybrid cloud management system that dynamically allocates WOD resources for storage access traffic toward a storage system according to the protocol type of the traffic.
[0065] FIG. 21 illustrates an example of a network that contains a cloud network manager 2120 which has a capability to allocate WOD resources dynamically, in accordance with an example implementation. Storage manager 2122 and cloud network manager 2120 are connected by LAN 2121 and router 2123 to the WAN 2140. The cloud network manager 2120 also has a capability to detect the traffic types and selects appropriate WAN optimization methods. The WODs 2109 and 2118 detect storage access packets and notify the cloud network manager 2120 via routers 2139 and 2119 which are connected to WAN 2140 which connects to router 2123 and LAN 2121. The cloud network manager 2120 then configures the WOD 2109 and 2118. The storage access has three types; iSCSI READ from the host 2110 for the data that is stored on the storage system 2103, iSCSI READ from the storage controller 2114 of the storage system 2113, iSCSI WRITE from the controller 2104 of the storage system 2103 to the volume 2116 on the storage system 2113. Storage system 2113 has an additional volume 2115 and storage system 2103 has volumes 2105-2107. Host 2101 contains VSW 2102 and connects to LAN 2108. Host 2110 contains VM 2111 and VSW 2112 and connects to LAN 2117. [0066] FIG. 22 illustrates an example of the unknown flow information composed of the flow information 2201 that is sent from the WODs 2109 and 2118 and is received by the cloud network manager 2120, in accordance with an example implementation. The three entries represent the three types of iSCSI traffic which are described above.
[0067] FIG. 23 illustrates an example of the flow information 2301 of the cloud network manager 2120, in accordance with an example implementation. Flow information 2301 has an extended structure compared with the flow information 409 described in the FIG. 15. Flow information 2301 has a "Optimized Method" column in its table. The column can take "Cache" or "TCP" as its value. "Cache" indicates that the WOD corresponding to the entry uses caching method for the specified flow. "TCP" indicates that the WOD corresponding to the entry uses TCP-window control method for the specified flow.
[0068] FIG. 24 illustrates an example of the replication information 2401 which the storage manager 2122 stores to manage the configurations of volume replication between two hosts, in accordance with an example implementation. Replication information 2401 is implemented as a table wherein each entry is composed of a destination storage system ID, a destination volume ID, an ID of the destination-side controller used for the replication, source storage system ID, a source volume ID, and an ID of the source-side controller used for the replication.
[0069] FIG. 25 illustrates an example of the flow chart of the set-up process of the cloud network manager 2120, in accordance with an example implementation. At first the cloud network manager 2120 receives an unknown flow packet at 2601 and checks at 2602 if the unknown (unregistered) flow is an iSCSI READ flow. If it is, the cloud network manager 2120 retrieves iSCSI initiator/target IP address and logical unit (LUN) information at 2603, retrieves volume migration information 1421 at 2604 and checks if the iSCSI READ flow is designated to a migrated volume at 2605. If the iSCSI READ flow is designated to the migrated volume (Y), the cloud network manager 2120 utilizes a TCP-window control method to modify the flow across the WAN. Otherwise (N), the cloud network manager checks for an iSCSI WRITE check flow at 2608, as described in FIG. 26.
[0070] The TCP-window control method is utilized so that once a data block stored on a storage system 2103 is read by the storage system 2113 the data block is stored on the storage system 2113 and is not read across the WAN. The caching method is therefore not needed in this instance. Finally if the volume is migrated at 2605 (Y) the cloud network manager 2120 registers the opposite flow to target flow information 2301 at 2606. The opposite flow means the iSCSI READ response replied from the storage system 2103 to the storage system 2113.
[0071] On the other hand, if the iSCSI flow is for the regular volume and not for the migrated volume (N) at 2605, the cloud network manager 2120 utilizes a caching method and stores the read data block on the WOD 2118 at 2607. [0072] Further, if the flow is iSCSI WRITE from the storage system 2103 to the storage system 2113, the cloud network manager 2120 processes an iSCSI WRITE check flow at 2608.
[0073] FIG. 26 illustrates an example of the iSCSI WRITE check flow, in accordance with an example implementation. At first, the cloud network manager 2120 checks if the unknown flow is an iSCSI WRITE Flow or not at 2609. If it is (Y), the cloud network manager 2120 retrieves the corresponding volume for the flow from the replication information 2401 at 2610. If a corresponding entry is found at 2611 (Y), the cloud network manager 2120 registers the flow as a flow that is optimized with the TCP-windows control method at 2612. If not, the flow is registered as a flow that is optimized with the caching method at 2613.
[0074] By using the above flow, especially the flow that the cloud network manager 2120 distinguishes the types of flows among hosts and storage systems distributed across the WAN 2140, the example implementation can allow users to utilize limited resources of WAN optimization devices efficiently and to keep the performance of iSCSI traffic flows for storage volume migration.
[0075] Moreover, other implementations of the present application will be apparent to those skilled in the art from consideration of the specification and practice of the teachings of the present application. Various aspects and/or components of the described example implementations may be used singly or in any combination. It is intended that the specification and example implementations be considered as examples only, with the true scope and spirit of the present application being indicated by the following claims.