Computer system having plural of storage systems

BACKGROUND OF THE INVENTION

P-0001[0001] The present invention relates to a computer system virtually managing or controlling storage areas of a plurality of storage systems as one logical storage area or a plurality of logical storage areas.

P-0002[0002] Recently, the amount of data used in an information processing system has been considerably increased and hence storage management expenses imposed on the user have been increased. To cope with this problem, a report “Virtualizing the SAN” published on Jul. 5, 2000, pp. 1-20 from an investigating company called “Morgan Keegan” describes a system including a plurality of storage systems. In the system, storage areas of the storage systems are virtualized into logical storage areas (virtual volumes). By allocating the virtual volumes to users (computers), the storage management expenses of the system users are reduced. The report describes two approaches on a system configuration and a method to implement the virtualization.

P-0003[0003] In this specification, a storage system is used to indicate a single storage such as a disk device and a set of “just bunch of disks (JBOD)” and a controller thereof as well as a set of a plurality of storages such as redundant array of inexpensive disks (RAID) and a controller thereof.

P-0004[0004] The first approach provides a virtualization system in which a computer connected between a computer (to be called a server hereinbelow) using the storage system and the storage system manages or controls a correspondence between storage areas of the storage system and virtual volumes indicated to the server. In the system, having received an access request to a virtual volume, the computer converts the request to an access request to a storage area of the storage system and issues the request to the storage system.

P-0005[0005] The second approach provides a virtualization system including a virtualization or virtualizing computer, which controls a correspondence between storage areas of the storage system and an area of a virtual volume indicated to the user. In the system, to access a virtual volume, a server issues an inquiry to the virtualization computer to obtain a position or an address of a storage area of a storage system corresponding to the virtual volume to be accessed. In response to the inquiry, the virtualization computer sends an address in the storage area to the server according to information managed by the virtualization computer. Having received the address from the virtualization computer, the server accesses the area of the storage system using the address.

P-0006[0006] According to the second approach, the server must collaborate with the virtualization computer. In consequence, it is necessary to prepare a collaborative function for the virtualization computer in association with many types of operating systems (OS) of servers. However, considerably many types of general operating systems are available in the present stage of art. It is therefore not practical to cope with all operating systems.

P-0007[0007] On the other hand, when compared with the second approach, the first approach is advantageous because it is not required to cope with the respective operating systems. However, the computer connected to the servers and the storage systems executes processing such as the address translation and the issuance of a request to a storage system. This leads to a problem that a limit of performance of the computer becomes a limit of performance of the system. For example, saving data from a storage system to another storage system imposes quite a heavy load on the computer. In this case, access performance of the overall system is also reduced. If the computer implementing the virtualization fails, the entire storage system cannot be used.

SUMMARY OF THE INVENTION

P-0008[0008] It is therefore an object of the present invention to provide a virtualization system capable of mitigating the load imposed on an apparatus (virtualization apparatus) to perform a virtualization processing between a server and a storage system.

P-0009[0009] Another object of the present invention is to provide a virtualization system capable of achieving desired performance in association with increase in the load imposed on the virtualization apparatus executing virtualization processing between a server and a storage system.

P-0010[0010] Further another object of the present invention is to provide a virtualization system in which even if the computer managing virtualization fails, input and output operations of data are not adversely influenced between a server and a storage system.

P-0011[0011] To achieve the objects according to the present invention, the virtualization apparatus to conduct virtualization includes a switch (to be referred to as a virtualization switch hereinbelow) to execute virtualization processing and a computer (to be referred to as a virtualization server) to execute virtualization processing. In the configuration, the virtualization switch executes processing as follows. The switch receives data from the server to convert the data into data to be fed to a storage area of the storage system and transmits the data to the storage system. Or, the virtualization switch converts the received data into data from a virtualized volume and sends the data to the server. In the operation, the virtualization apparatus uses virtualization information, i.e., information of a correspondence between virtual volumes and storage areas of the storage system.

P-0012[0012] On the other hand, the virtualization server receives, from the virtualization switch, data requiring predetermined processing, for example, data to be copied from a storage system onto another storage system. The server executes the predetermined processing and then sends the data to the server or an associated storage device.

P-0013[0013] In the configuration in which the virtualization device includes a plurality of virtualization servers and a plurality of virtualization switches, when one of the switches or the servers fails, it is possible that the processing of virtualization is taken over to, for example, another virtualization server.

P-0014[0014] A configuration of a favorable embodiment of the present invention includes a computer for management (control server) which controls virtual information indicating a correspondence between storage areas of the storage system and virtual volumes and which also controls states of devices in the virtualization system. In the configuration, when the control server issues an instruction to the virtualization apparatus to change an access path between the server and the storage system, the virtualization apparatus allocates several ones of the virtual volumes as processing objects in a section of the virtualization apparatus to processing of devices with a lower processing load.

P-0015[0015] According to the present invention, there is provided a virtualization system in which the computer to conduct the virtualization does not become a processing bottleneck in the overall system. Moreover, according to the present invention, there is provided a virtualization system which achieves desired performance in association with increase in the processing load in a section executing the virtualization processing.

P-0016[0016] According to the present invention, there is provided a virtualization system in which even when a failure occurs in the virtualization apparatus, the failure does not exert adverse influence on any external device.

P-0017[0017] Other objects, features and advantages of the invention will become apparent from the following description of the embodiments of the invention taken in conjunction with the accompanying drawings.

BRIEF DESCRIPTION OF THE DRAWINGS

P-0018[0018]FIG. 1 is a diagram showing a configuration of an embodiment of a system according to the present invention.

P-0019[0019]FIG. 2 is a diagram showing programs and tables stored in a memory of a control server.

P-0020[0020]FIG. 3 is a diagram showing programs and tables stored in a memory of a virtualization server.

P-0021[0021]FIG. 4 is a diagram showing programs and tables stored in a memory of a virtualization switch.

P-0022[0022]FIG. 5 is a diagram showing an example of a virtual volume control table in the control server.

P-0023[0023]FIG. 6 is a diagram showing an example of a real volume control table in the control server.

P-0024[0024]FIG. 7 is a diagram showing an example of a state control table of the virtualization switch in the control server.

P-0025[0025]FIG. 8 is a diagram showing an example of a state control table of the virtualization server in the control server.

P-0026[0026]FIG. 9 is a diagram showing an example of a virtual volume control table in the virtualization server.

P-0027[0027]FIG. 10 is a diagram showing an example of a virtual volume control table in the virtualization switch.

P-0028[0028]FIG. 11 is a flowchart showing a processing flow of a volume control program of the control server.

P-0029[0029]FIG. 12 is a flowchart showing a processing flow of a device state control program of the control server.

P-0030[0030]FIG. 13 is a flowchart showing a processing flow of a device state control program of the virtualization server.

P-0031[0031]FIG. 14 is a flowchart showing a flow of virtualization processing of the virtualization switch.

P-0032[0032]FIG. 15 is a flowchart showing a flow of virtualization processing of the virtualization server.

P-0033[0033]FIG. 16 is an example of commands exchanged between a virtualization node and a control server.

DESCRIPTION OF THE EMBODIMENTS

P-0034[0034]FIG. 1 shows a configuration of an embodiment of a system according to the present invention. The system includes servers <b>101</b>, storage systems <b>106</b>, a virtualization apparatus (virtualization node) <b>103</b> connected between the servers <b>101</b> and the storages <b>106</b>, and a control server <b>102</b>.

P-0035[0035] The servers <b>101</b> are connected via networks <b>110</b> to the virtualization node <b>103</b>. The storage systems <b>106</b> are connected via a network <b>111</b> to the virtualization node <b>103</b>. The networks <b>110</b> and <b>111</b> may be networks of one type of protocols or networks of mutually different types of protocols. For example, it is possible that the networks <b>110</b> are networks of transmission control protocol/internet protocol (TCP/IP) and the networks <b>111</b> are fiber-channel networks.

P-0036[0036] Moreover, the networks <b>110</b> may be of the same type for each server connected thereto or may be of mutually different types for the respective servers. For example, networks of TCP/IP and networks of the fiber-channel protocol can be used. Additionally, the networks <b>111</b> may also be of the same type for each storage system connected thereto or may be of mutually different types for the respective storage systems.

P-0037[0037] The virtualization node <b>103</b> supplies to the server <b>101</b> a virtual volume configured by virtualizing storage areas of the storage system <b>106</b>. The control server <b>102</b> manages a correspondence between the storage areas of the storage system <b>106</b> and the storage areas and controls the virtualization node <b>103</b>.

P-0038[0038] In this specification, a “storage area” is a storage area indicated by the storage system <b>106</b> to another device. However, the storage area indicated by the storage system <b>106</b> has a one-to-one correspondence to a physical storage area of a disk device <b>140</b> of the storage system <b>106</b> or matches a logical storage area virtualized in the storage system <b>106</b> such as a logical unit (LU) of an RAID system. In this specification, it is not particularly discriminated whether a “storage area” is a logical or physical storage area of the storage system <b>106</b>. In this connection, a storage area indicated by the storage system <b>106</b> will be referred to as a real volume hereinbelow.

P-0039[0039] The server <b>101</b> includes a central processing unit (CPU) <b>134</b> to execute programs accessing the storage systems <b>106</b>, ports <b>121</b> to conduct communication of data with the storage systems <b>106</b>, and a port <b>120</b> connected via a network <b>113</b> to the control server <b>102</b>. The server <b>101</b> includes at least one port <b>121</b>, but may include two or more ports <b>121</b>.

P-0040[0040] The storage system <b>106</b> includes ports <b>127</b> to communicate data with external devices, a port <b>128</b> connected via the network <b>113</b> to the control server <b>102</b>, disk devices <b>140</b> to store data therein, and a control unit <b>139</b> to operate a program controlling, for example, data communication with the disk devices <b>140</b>. The storage system <b>106</b> includes at least one port <b>127</b>, but may include two or more ports <b>127</b>.

P-0041[0041] The control server <b>102</b> includes a memory <b>132</b>, a storage <b>133</b>, a CPU <b>131</b> which executes programs to manage a correspondence between real volumes and virtual volumes of the storage systems <b>106</b> according to information stored in the storage <b>133</b> and which executes programs to manage a state of the virtualization node <b>103</b> to control the virtualization node <b>103</b> so that the node <b>103</b> provides stable volume accesses to the server <b>101</b>, and a port <b>129</b> to communicate data via the network <b>113</b> with external devices such as the server <b>101</b>, the virtualization node <b>103</b>, and the storage systems <b>106</b>.

P-0042[0042] The virtualization node <b>103</b> includes a virtualization switch <b>105</b> and a virtualization server <b>104</b>. However, the virtualization node <b>103</b> may be in a configuration in which the switch <b>105</b> and server <b>104</b> are integrally configured as one unit or are connected to each other via a network or the like. Additionally, the node <b>103</b> desirably includes a plurality of virtualization switches <b>105</b> and a plurality of virtualization servers <b>104</b>.

P-0043[0043] The virtualization switch <b>105</b> receives data sent from the server <b>101</b> to a virtual volume, converts the data into data to be sent to a real volume, and sends the data to the storage system <b>106</b>. Moreover, the switch <b>105</b> converts data received from the storage system into data to be sent to a virtual volume and sends the data to the server <b>101</b>.

P-0044[0044] The switch <b>105</b> includes port <b>122</b> to communicate data with the servers <b>101</b>, ports <b>124</b> to communicate data with the virtualization servers <b>104</b>, ports <b>123</b> to communicate data with the storage system <b>106</b>, a port <b>141</b> to communicate data with another virtualization switch <b>105</b>, a port <b>130</b> to communicate data via the network <b>113</b> with the control server <b>102</b>, a memory <b>136</b>, and a CPU <b>135</b> to execute a program which converts data received from a server <b>101</b> according to information stored in the memory <b>136</b> to send the data to the storage systems <b>106</b> and which converts data received from a storage system <b>106</b> to send the data to an associated server <b>101</b>.

P-0045[0045] However, the configuration may include one or more ports <b>141</b> to communicate data between the switches <b>105</b>. Or, the switch <b>105</b> may be omitted depending on cases.

P-0046[0046] Furthermore, when the virtualization node <b>103</b> includes a plurality of virtualization switches <b>105</b>, the servers <b>101</b> are connected to the switches <b>105</b> in one of two ways as follows.

P-0047[0047] First, each server <b>101</b> is connected to a plurality of switches. Second, a plurality of servers <b>101</b> are connected via ports <b>141</b> to each other such that either one of the switches <b>105</b> is connected to the servers <b>101</b>.

P-0048[0048] In either cases, when a server <b>101</b> detects a failure in one of the virtualization switches <b>105</b> or when a server <b>101</b> receives information of a failure in a switch <b>105</b> from a control server <b>102</b> or the like, the server <b>101</b> changes an access route to the virtualization node <b>103</b>, specifically, changes the switch <b>105</b>.

P-0049[0049] Incidentally, the configuration may include at least one port <b>122</b> to communicate data with the servers <b>101</b>, at least one port <b>124</b> to communicate data with the virtualization servers <b>104</b>, and at least one port <b>123</b> to communicate data with the storage systems <b>106</b>.

P-0050[0050] The virtualization server <b>104</b> receives data requiring predetermined processing from the virtualization switch <b>105</b>, the data being selected from data beforehand received by the switch <b>105</b>. The server <b>104</b> then executes the processing. Thereafter, the virtualization server <b>104</b> sends the processed data via the switch <b>105</b> to the server <b>101</b> or the storage system <b>106</b>.

P-0051[0051] The virtualization server <b>104</b> includes ports <b>125</b> to communicate data via a network <b>114</b> with the virtualization switches <b>105</b>, a port <b>126</b> to communicate data via the network <b>113</b> with the control server <b>102</b>, a memory <b>138</b>, and a CPU <b>137</b> to execute processing such as remote copy processing according to predetermined information. Incidentally, the configuration includes at least one port <b>125</b> to be connected to the virtualization switch <b>104</b>.

P-0052[0052]FIG. 2 shows a layout of programs and tables of the control server <b>102</b>. The memory <b>132</b> stores a volume control program <b>201</b> to control information indicating a correspondence between real volumes and virtual volumes of the storage systems <b>106</b>, a device state control program <b>202</b> which controls a state of the virtualization node <b>103</b> to supply stable volume accesses to volumes to the servers <b>101</b>, a virtual volume control table <b>203</b>, a real volume control table <b>204</b>, virtualization switch state control table <b>205</b>, and a virtualization server state control table <b>206</b>.

P-0053[0053]FIG. 3 shows a layout of programs and tables of the virtualization server <b>104</b>. The memory <b>138</b> stores a volume control program <b>301</b>, a virtualization processing program <b>302</b> which receives data from the virtualization switches to execute predetermined processing such as data copy or replication between the storage systems, a device state control program <b>303</b> to monitor a state of the load imposed on each virtualization node <b>103</b>, a virtual volume control table <b>304</b>, and a device state control table <b>305</b>.

P-0054[0054] The volume control program <b>301</b> receives an instruction from the volume control program <b>201</b> of the control server <b>102</b> and is then executed by the virtualization server <b>104</b> to update the virtual volume control table <b>304</b>. Having received an enquiry from the control server <b>102</b>, the device state control program <b>303</b> is executed by the virtualization server <b>104</b> to notify a state of the virtualization server <b>104</b> at the point of enquiry, specifically, a utilization ratio or the like of the server <b>104</b>.

P-0055[0055] The device state control table <b>305</b> stores information of the state of the load on the virtualization server <b>104</b>, for example, a numeric value of the operation ratio or the like. The information of the load state may be, in addition to a numeric value, information of a degree, for example, “high” or “low”.

P-0056[0056]FIG. 4 shows a layout of programs and tables of the virtualization switch <b>105</b>. The memory <b>136</b> stores a volume control program <b>401</b>, a virtualization program to execute virtualization processing, a device state control program <b>403</b> to monitor a state of each device, a virtual volume control table <b>404</b>, and a device state control table <b>405</b>.

P-0057[0057] When an instruction is received from the volume control program <b>201</b> of the control server <b>102</b>, the volume control program <b>401</b> is executed by the virtualization switch <b>105</b> to update the virtual volume control table <b>404</b>. The virtualization processing executed by the virtualization program <b>402</b> in the switch <b>105</b> is as follows.

P-0058[0058] The virtualization switch <b>105</b> receives data sent from a server <b>101</b> to a virtual volume and converts the data into data for a real volume of an associated storage system <b>106</b> to send the data to the storage system <b>106</b>. Or, the switch <b>105</b> converts data received from the storage system <b>106</b> into data from a virtualized volume to send the data to the server <b>101</b>. In this connection, the conversion of data specifically means a conversion from a virtual volume address into a real volume address or a conversion from a real volume address into a virtual volume address.

P-0059[0059] When an enquiry request is received from the control server <b>102</b>, the device state control program <b>403</b> is executed by the virtualization switch <b>105</b> to notify a state of the switch <b>105</b> at the point of enquiry, specifically, a state of the input/output load on the switch or the like.

P-0060[0060] The device state control table <b>405</b> stores information indicating a state of the load on the switch <b>105</b>, for example, a numeric value indicating the number of input/output operations of the switch <b>105</b>. Incidentally, the information of the load state may be other than a numeric value, namely, may be information indicating a degree, for example, “high” or “low”.

P-0061[0061]FIG. 5 shows an example of the virtual volume control table in the control server <b>102</b>. In the example, an identifier (ID) of a port to access a virtual volume is registered to an entry <b>501</b>. Information indicating the virtual volume is registered to an entry <b>502</b>. An entry <b>502</b> includes an entry <b>507</b> to which a logical unit (LU) number of the virtual volume is registered and an entry <b>508</b> to which a size of the virtual volume is registered.

P-0062[0062] An identifier (ID) of the virtualization switch <b>104</b> controlling the pertinent virtual volume is registered to an entry <b>503</b>. An identifier (ID) of the virtualization server <b>104</b> which collaborates with the virtualization switch <b>105</b> to execute processing for the pertinent virtual volume is registered to an entry <b>504</b>. An identifier (ID) of a server <b>101</b> using the pertinent virtual volume is registered to an entry <b>505</b>. Information of real volumes constituting the pertinent virtual volume is registered to an entry <b>506</b>.

P-0063[0063] The entry <b>506</b> includes an entry <b>510</b> to which an identifier (ID) of a port to access a real volume is registered, an entry <b>511</b> to which a logic unit (LU) number of the real volume is registered, an entry <b>512</b> to which a size of the real volume is registered, and entries <b>513</b> and <b>514</b> to which a start address and an end address respectively indicating the start and end positions of an area of the real volume allocated to the pertinent virtual volume are respectively registered.

P-0064[0064] In this regard, the respective entries, excepting the entry <b>509</b>, of the constituent real volume information are prepared for each virtual volume as many as there are real volumes allocated to the virtual volume.

P-0065[0065] For example, in the table of FIG. 5, a first row indicates that virtual volume “0” having a virtual port ID of “123” and a size of one megabyte (MB) is controlled by virtualization switch “XXX” and virtualization server “DDD” and the virtual volume is associated with an area ranging from a first address of “8” to an end address of “1954” of one real volume, i.e., volume “8” having a port ID of “AAA” and a size of 1 MB.

P-0066[0066] A second row of the table indicates that virtual volume “1” having a virtual port ID of “123” and a size of one megabyte (MB) is controlled by virtualization switch “XXX” and virtualization server “DDD” and the virtual volume includes two areas, namely, an area ranging from first address “0” to end address “977” of real volume “8” having a port ID of “BBB” and a size of 500 kilobytes (KB) and an area ranging from first address “0” to end address “977” of real volume “3” having a port ID of “CCC” and a size of 500 KB.

P-0067[0067] In this case, when it is inhibited that real volumes are divided to be allocated to virtual volumes, the range of each real volume is not required and hence the entries <b>512</b> and <b>513</b> may be omitted. When the real volumes are allocated to the virtual volumes according to a one-to-one correspondence, it is only necessary to specify a real volume to each virtual volume and hence the entries <b>509</b>, <b>512</b>, <b>513</b>, and <b>514</b> may be omitted.

P-0068[0068]FIG. 6 shows an example of the real volume control table <b>204</b> in the control server <b>102</b>. Information indicating a storage system is registered to an entry <b>601</b>. A port identifier (ID) to access a real volume is registered to an entry <b>602</b>. A logical unit (LU) number of the real volume is registered to an entry <b>603</b>. An area of a real volume allocated to the virtual volume is registered to an entry <b>604</b>. Information indicating an area not used for the virtual volume, i.e., an available area is registered to an entry <b>605</b>.

P-0069[0069] Incidentally, when it is inhibited that real volumes are divided to be allocated to virtual volumes, the entries <b>604</b> and <b>605</b> may be omitted.

P-0070[0070]FIG. 7 shows an example of the state control table <b>205</b> in the virtualization switch <b>105</b> of the control server <b>102</b>. An identifier (ID) of a virtualization switch <b>105</b> is registered to an entry <b>701</b>. Information indicating a state, specifically, a “normal” or “abnormal” state of the switch <b>105</b> is registered to an entry <b>702</b>. In this case, “normal” indicates that the switch <b>105</b> conducts a predetermined operation and “abnormal” indicates that a failure has occurred in the switch <b>105</b>. Information indicating a load imposed on the switch <b>105</b> is registered to an entry <b>703</b>.

P-0071[0071]FIG. 8 shows an example of the state control table <b>206</b> in the virtualization server <b>101</b> of the control server <b>102</b>. An identifier (ID) of a virtualization server <b>101</b> is registered to an entry <b>801</b>. Information indicating a state, specifically, a “normal” or “abnormal” state of the server <b>101</b> is registered to an entry <b>802</b>. Information indicating a load imposed on the server <b>101</b> is registered to an entry <b>803</b>.

P-0072[0072]FIG. 9 shows an example of the virtual volume control table <b>304</b> of the virtualization server <b>101</b>. An identifier (ID) of a port to access a virtual volume is registered to an entry <b>901</b>. Information indicating the virtual volume is registered to an entry <b>902</b>. The entry <b>902</b> includes an entry <b>905</b> to which a logical unit (LU) number of the virtual volume is registered and an entry <b>906</b> to which a size of the virtual volume is registered.

P-0073[0073] An identifier (ID) of a virtualization switch <b>105</b> controlling the pertinent virtual volume is registered to an entry <b>903</b>. Information indicating the virtual volume is registered to an entry <b>904</b>. Specifically, the entry <b>904</b> includes an entry <b>907</b> to which information indicating the number of real volumes allocated to the virtual volume is registered, an entry <b>908</b> to which an identifier (ID) of a port to access a virtual volume is registered, an entry <b>909</b> to which a logic unit (LU) number of the virtual volume is registered, an entry <b>910</b> to which a size of the real volume is registered, and entries <b>911</b> and <b>912</b> to which a start address and an end address respectively indicating the start and end positions of an area of the real volume allocated to the pertinent virtual volume are respectively registered.

P-0074[0074] Incidentally, the respective entries, excepting the entry <b>907</b>, of the entry <b>904</b> are prepared for each virtual volume as many as there are real volumes allocated to the virtual volume.

P-0075[0075] When it is inhibited that real volumes are divided to be allocated to virtual volumes, the entries <b>911</b> and <b>912</b> may be omitted. When the real volumes are allocated to the virtual volumes according to a one-to-one correspondence, the entries <b>907</b>, <b>910</b>, <b>911</b>, and <b>912</b> may be omitted.

P-0076[0076]FIG. 10 shows an example of the virtual volume control table <b>404</b> of the virtualization switch <b>105</b>. An identifier (ID) of a port for the server <b>101</b> to access a virtual volume is registered to an entry <b>1001</b>. Information about the virtual volume is registered to an entry <b>1002</b>. An entry <b>1002</b> includes an entry <b>1005</b> to which a logical unit (LU) number of the virtual volume is registered and an entry <b>1006</b> to which a size of the virtual volume is registered.

P-0077[0077] An identifier (ID) of a virtualization server <b>104</b> controlling the pertinent virtual volume is registered to an entry <b>1003</b>. Information indicating information of real volumes constituting the virtual volume is registered to an entry <b>1004</b>. Specifically, the entry <b>1004</b> includes an entry <b>1007</b> to which the number of real volumes is registered, an entry <b>1008</b> to which an identifier (ID) of a port to for the server to access a real volume is registered, an entry <b>1009</b> to which a logic unit (LU) number of the real volume is registered, an entry <b>1010</b> to which a size of the real volume is registered, and entries <b>1011</b> and <b>1012</b> to which a start address and an end address respectively indicating the start and end positions of an area of the real volume allocated to the pertinent virtual volume are respectively registered.

P-0078[0078] In this regard, the respective entries, excepting the entry <b>1007</b>, of the entry <b>1004</b> are prepared for each virtual volume as many as there are real volumes allocated to the virtual volume.

P-0079[0079] When it is inhibited that real volumes are divided to be allocated to virtual volumes, the entries <b>1002</b> and <b>1003</b> may be omitted. When the real volumes are allocated to the virtual volumes according to a one-to-one correspondence, the entries <b>1007</b>, <b>1010</b>, <b>1011</b>, and <b>1012</b> may be omitted.

P-0080[0080] The control server <b>102</b> executes processing in response to an instruction from the user as follows. According to a quantity of storage capacity desired by the user, the server <b>102</b> selects an appropriate number of real volumes from the real volumes of the storage systems <b>106</b> to set the selected real volumes as a virtual volume and then notifies a result of the operation to the user. For this processing, the server <b>102</b> executes the volume control program <b>201</b>.

P-0081[0081]FIG. 11 shows in a flowchart a flow of processing of the volume control program <b>201</b> executed by the control server <b>102</b>.

P-0082[0082] The server <b>102</b> executes processing such as initialization to read the control tables of the control server <b>102</b> (<b>1101</b>). The server <b>102</b> judges to determine whether or not an indication from the user is a virtual volume allocation request (<b>1102</b>). If the indication is a virtual volume allocation request, the server <b>102</b> judges to determine, by using the real volume control table <b>204</b>, whether or not the storage systems <b>106</b> have available real volumes (<b>1105</b>).

P-0083[0083] If there exist available real volumes, the control server <b>102</b> extracts therefrom a storage area of storage capacity requested by the user and defines the area as one virtual volume or a plurality of virtual volumes. The server <b>102</b> then updates the contents of the real volume control table <b>204</b> and the virtual volume control table <b>203</b>. Specifically, the contents of the tables are changed such that information indicating the defined virtual volumes is assigned to the available real volumes (<b>1107</b>).

P-0084[0084] Thereafter, the control server <b>102</b> notifies the allocation of the real volumes to the storage systems <b>106</b> (<b>1108</b>). The server <b>102</b> instructs the virtualization node <b>103</b>, specifically, the virtualization server <b>104</b> and the virtualization switch <b>105</b> to update the virtual volume control table (<b>1109</b>). According to the request of the user, the server <b>102</b> notifies the allocation of the available volumes to the virtual volumes (<b>1110</b>). The server <b>102</b> then returns to step <b>1102</b> and waits for an indication from the user.

P-0085[0085] If no available volume is detected in step <b>1105</b>, the control server <b>102</b> indicates to the user the impossible allocation of virtual volumes requested by the user (<b>1106</b>) and returns to step <b>1102</b> to wait for an indication from the user.

P-0086[0086] If the user request is other than a virtual volume allocation request in step <b>1102</b>, the control server judges to determine whether or not the user request is a request to terminate the processing (<b>1103</b>). If the request is a processing termination request, the server <b>102</b> terminates the processing. Otherwise, the server <b>102</b> executes processing associated with the user request (<b>1104</b>) and returns to step <b>1102</b> to wait for an indication from the user.

P-0087[0087] Incidentally, the user indication may be directly inputted to the control server <b>102</b> or may be inputted from another program using a command. Additionally, the notification from the control server <b>102</b> to the user may be presented on a display of the control server <b>102</b> or may be outputted by another program using a command. The step to notify the real volume allocation to the storage systems <b>106</b> may be omitted.

P-0088[0088] The control server <b>102</b> monitors the state of the virtualization node <b>103</b>. When a failure or a high load appears in a switch <b>105</b> or the like of the virtualization node, the server <b>102</b> issues an indication to the virtualization node or the server <b>101</b> to change an access route from the server <b>101</b> of the user to the storage system <b>106</b> to avoid a failure or the like. Through the operation, overall system performance can be kept unchanged. The processing is achieved when the server <b>102</b> executes the device state control program <b>202</b>. The program <b>202</b> is repeatedly executed at a particular interval of time during the system operation or is executed in response to an indication from the user.

P-0089[0089]FIG. 12 shows a flow of processing of the device state control program <b>202</b> executed by the control server <b>102</b>. The server <b>102</b> executes processing such as initialization to read the control tables (<b>1201</b>). The server <b>102</b> compares information of a predetermined time with the present time to determine whether or not the present time corresponds to timing to confirm the device state (<b>1202</b>).

P-0090[0090] If the present time corresponds to timing to confirm the device state, the control server <b>102</b> accesses the virtualization switch state control table <b>205</b> and the virtualization server state control table <b>206</b> to issue an enquiry of device state to the virtualization switches <b>105</b> and the virtualization servers <b>104</b>.

P-0091[0091]FIG. 16 shows an example of enquiry from the control server to the virtualization node and an example of a reply command to the enquiry. In the enquiry operation, information indicating presence or absence of a device failure and information of a state of the load imposed on the device are communicated. When the virtualization node <b>103</b> does not reply to an enquiry from the control server <b>102</b> to the virtualization node <b>103</b>, the server <b>102</b> assumes that a failure occurs in the node <b>103</b> (<b>1203</b>).

P-0092[0092] When a failure is detected in a device as a result of the enquiry (<b>1204</b>), the control server <b>102</b> replaces the access route passing through the device in which the failure is detected with an access route passing through a device in which any failure is not detected. Specifically, the server <b>102</b> accesses the virtual volume control table <b>203</b> to retrieve virtual volumes supplied to the server <b>101</b> via the virtualization server <b>104</b> or the virtualization switch <b>105</b> in which the failure is detected. Thereafter, the server <b>102</b> changes the access routes so that the retrieved virtual volumes are supplied to the server via a normal virtualization switch <b>105</b> or a normal virtualization server <b>104</b>. The server <b>102</b> accordingly updates the virtual volume control table (<b>1205</b>).

P-0093[0093] The control server <b>102</b> indicates the virtualization server <b>104</b> and the virtualization switch <b>105</b> to update the virtual volume control table. The server <b>102</b> also transmits information of the device in which the failure has occurred to the server <b>101</b> according to necessity (<b>1206</b>). The server <b>102</b> then updates the virtualization switch state control table <b>205</b> and the virtualization server state control table <b>206</b>. Specifically, the entry corresponding to the failed device is changed from “normal” to “abnormal” (<b>1207</b>).

P-0094[0094] After the processing is finished, the control server <b>102</b> returns to step <b>1202</b> and judges to determine whether or not the present time is timing to confirm the device state. If no failure is detected in step <b>1204</b>, the sever <b>102</b> confirms the load state of the virtualization node <b>103</b>. Specifically, the server <b>102</b> confirms the load state obtained as a result of the enquiry to the device (<b>1208</b>).

P-0095[0095] When it is detected that the load of the virtualization switch <b>105</b> or the virtualization node <b>104</b> in the virtualization node <b>103</b> exceeds a predetermined value, e.g., 70%, the control server <b>102</b> changes the access route not to use the device of which the load exceeds the predetermined value, namely, the server <b>102</b> takes a detour. Specifically, the server <b>102</b> accesses the virtual volume control table <b>203</b> to retrieve processing associated with the virtual volumes supplied to the server <b>101</b> via the virtualization switch <b>105</b> or the virtualization node <b>104</b> of which the load exceeds the predetermined value. The server <b>102</b> selects particular processing from the retrieved processing and changes access routes of selected processing so that the selected processing is supplied via another virtualization switch <b>105</b> or another virtualization node <b>104</b>. The server <b>102</b> updates the virtual volume control table <b>203</b>.

P-0096[0096] Thereafter, the control server <b>102</b> indicates the virtualization switch <b>105</b> and the virtualization node <b>104</b> to update the virtualization volume control table. According to necessity, the server <b>102</b> sends information of the device of which the load is equal to or more than the predetermined value. The virtualization node <b>103</b> having received the indication updates the virtualization volume control table. This establishes the new access routes (<b>1210</b>). The server <b>102</b> updates the virtualization switch state control table <b>205</b> and the virtualization server state control table <b>206</b> (<b>1207</b>).

P-0097[0097] The control server <b>102</b> returns to step <b>1202</b> to determine whether or not the present time corresponds to timing to confirm the device state. If the load is less than the predetermined value in step <b>1208</b>, the server <b>102</b> updates, according to the information received from the virtualization node <b>103</b>, the virtualization switch state control table <b>205</b> and the virtualization server state control table <b>206</b> in step <b>1207</b> and then returns to step <b>1202</b>.

P-0098[0098] If it is not determined in step <b>1202</b> that the present time corresponds to timing to confirm the device state, the control server <b>102</b> judges to determine whether or not a device state notification has been received from the virtualization switch <b>105</b> or the virtualization server <b>104</b>. Specifically, the server <b>102</b> confirms whether or not a communication control section, not shown, has received a device state notification from each device. Through the confirmation, even if a failure occurs at timing other than the periodic state confirmation point, it is possible to appropriately change associated access routes (<b>1211</b>). If the notification has been received, the server <b>102</b> executes processing of step <b>1204</b> and subsequent steps.

P-0099[0099] If the device state notification has not been received and an end indication has not been notified from the user in step <b>1211</b>, the control server <b>102</b> returns to step <b>1202</b>. If the end indication has been received in step <b>1212</b>, the server <b>102</b> terminates the processing. Specifically, when the user indicates the end of processing or when the server <b>102</b> recognizes that the predetermined time is past, the server <b>102</b> assumes that the end indication is received and terminates the processing.

P-0100[0100] On the other hand, having received an indication from the control server <b>102</b> to update the virtual volume control table, the virtualization server <b>104</b> updates the control table using the volume control program <b>301</b>.

P-0101[0101] Specifically, the virtualization server <b>104</b> confirms by the volume control program <b>301</b> whether or not an indication to update the virtual volume control table has been received from the control server <b>102</b>. If it is confirmed that the reception of the update indication has been received from the server <b>102</b>, the virtualization server <b>104</b> updates the virtual volume control table <b>304</b> according to the update indication.

P-0102[0102] The update processing similarly applies to the virtualization switch <b>105</b>.

P-0103[0103] The virtualization switch <b>105</b> and the virtualization server <b>104</b> having updated the virtual volume control table <b>304</b> execute virtualization of the received data according to the contents of the updated virtual volume control table <b>304</b>.

P-0104[0104] Assume that the control server <b>102</b> has detected an event that a failure occurs in a virtualization server <b>104</b> or a load imposed thereon is equal to or more than a predetermined value. In this situation, for example, if a failure has occurred in the virtualization server <b>104</b>, the control server <b>102</b> updates the virtual volume control table <b>203</b> to form an access route not using the virtualization server <b>104</b>. If the load imposed thereon is equal to or more than a predetermined value, the control server <b>102</b> updates the virtual volume control table <b>203</b> to form an access route such that part of the virtual volume processing being executed by the virtualization server <b>104</b> is passed to another virtualization server having a lower load. Thereafter, the control server <b>102</b> indicates the virtualization switch <b>105</b> and the virtualization server <b>104</b> to update the virtual volume control table. Having received the update indication, the switch <b>105</b> updates the table according to the indication. The switch <b>105</b> processes data to avoid the failed virtualization server <b>104</b> according to the updated virtual volume control table. As a result, the data is smoothly processed.

P-0105[0105] When a device state enquiry is received from the control server <b>102</b>, the virtualization node <b>103</b> confirms its own state and notifies the state to the server <b>102</b>. Or, the virtualization node <b>103</b> periodically confirms its own state and notifies the state to the server <b>102</b>. The processing is achieved by executing the device state control program <b>303</b> and the like. The program <b>303</b> is executed according to an indication from the user of the control server when the virtualization node <b>103</b> is initiated and is repeatedly executed at a particular interval of time.

P-0106[0106]FIG. 13 shows processing of the device state control program <b>303</b> in the virtualization server <b>104</b>.

P-0107[0107] The server <b>104</b> conducts initialization (<b>1301</b>). The server <b>104</b> monitors its own state to update the device state control table <b>305</b> according to a change in the state (<b>1302</b>). The server <b>104</b> confirms whether or not a device state enquiry has been received from the control server <b>102</b> (<b>1303</b>). If the enquiry has been received, the server <b>104</b> notifies its own state controlled by the table <b>305</b> (<b>1304</b>).

P-0108[0108] If the enquiry has not been received in step <b>1303</b>, the virtualization server <b>104</b> judges to determine occurrence of a state to be notified to the control server <b>102</b>. For example, the server <b>104</b> determines whether or not the state is within a predetermined load (e.g., whether or not a load ratio of 70% is exceeded) or whether or not a failure has occurred in any section of the server <b>104</b> (<b>1305</b>).

P-0109[0109] If the state is to be notified to the control server <b>102</b>, the virtualization server <b>104</b> notifies its own state to the server <b>102</b> (<b>1304</b>). If the state is not to be notified in step <b>1305</b>, the virtualization server <b>104</b> judges to determine whether or not an end indication has been received (<b>1306</b>).

P-0110[0110] If such an end indication has been received, the server <b>104</b> terminates the program to finish the processing. Otherwise, the server <b>104</b> returns to step <b>1302</b>. In this connection, the processing of step <b>1305</b> may be omitted. The end indication may be, for example, an indication from the user or an indication from a timer in response to a lapse of a predetermined period of time.

P-0111[0111] The processing described in conjunction with FIG. 13 is almost the same as the processing achieved when the virtualization switch <b>105</b> executes the device state control program <b>403</b>.

P-0112[0112] In the virtualization node <b>103</b>, according to the contents of processing of data received from a server <b>101</b> or a storage system <b>106</b>, the virtualization switch <b>105</b> or the virtualization server <b>104</b> processes the data to send a result of the processing to the storage system <b>106</b> or the server <b>101</b>.

P-0113[0113]FIG. 14 shows a flow of the virtualization processing <b>402</b> executed by the virtualization switch <b>105</b>.

P-0114[0114] The switch <b>105</b> conducts initialization (<b>1401</b>). Having received data from a server <b>101</b> or a storage system <b>106</b> (<b>1402</b>), the switch <b>105</b> analyzes the data according to address information in the data.

P-0115[0115] Specifically, the switch <b>105</b> determines that the data is sent from a storage system (a real volume allocated to a virtual volume), the data is sent to a server <b>101</b> (a virtual volume), or the data source and destination are a storage system <b>106</b> or a server <b>101</b>. According to a result of the analysis, the switch <b>105</b> judges to determine whether or not the data requires virtualization (<b>1403</b>). If the data requires virtualization, the switch <b>105</b> judges to determine according to the virtual volume control table <b>404</b> whether or not the data is to be processed by the virtualization server <b>104</b>, for example, to be copied by the server <b>104</b> (<b>1404</b>).

P-0116[0116] If the data is to be processed by the virtualization server <b>104</b>, the virtualization switch <b>105</b> transfers the data to the server <b>104</b> (<b>1405</b>). Thereafter, the switch <b>105</b> returns to step <b>1402</b>.

P-0117[0117] If it is determined in step <b>1404</b> that the data does not require the processing of the virtualization server <b>104</b>, the switch <b>105</b> executes virtualization processing. The processing is specifically executed as below. If the data transmission source is a real volume allocated to a virtual volume, the switch <b>105</b> treats the data as data from a virtual volume. If the data transmission source is a virtual volume, the switch <b>105</b> treats the data as data from a real volume (<b>1406</b>).

P-0118[0118] Thereafter, the virtualization switch <b>105</b> transmits the processed data to the destination (<b>1407</b>) and executes processing of step <b>1402</b>.

P-0119[0119] If the data does not require the virtualization in step <b>1403</b>, the switch <b>105</b> transmits the data to the destination (<b>1407</b>) and executes processing of step <b>1402</b>. If it is determined in step <b>1402</b> that data has not been received, the switch <b>105</b> repeatedly executes processing of step <b>1402</b>.

P-0120[0120] In this regard, the judgement to determine whether or not the received data is to be processed by the virtualization server <b>104</b> or the like may be conducted without using the virtual volume control table <b>404</b>. The judgement may be conducted according to another aspect, for example, according to command information contained in the received data. It is also possible to omit step <b>1404</b>. In this case, the entire processing of the virtual volume is transferred to the virtualization server <b>104</b>.

P-0121[0121]FIG. 15 shows a flow of the virtualization processing <b>302</b> in the virtualization server <b>104</b>.

P-0122[0122] The server <b>104</b> conducts initialization (<b>1501</b>). When data is received from a virtualization switch <b>105</b> (<b>1502</b>), the server <b>104</b> judges to determine whether or not the data is to be processed by the virtualization server <b>104</b>, for example, whether or not the data is copied by the server <b>104</b> (<b>1503</b>). If the data is to be processed by the server <b>104</b>, the server executes an associated processing for the data.

P-0123[0123] Specific examples of processing executed by the virtualization server <b>104</b> are as follows. In data replication between storage systems <b>106</b>, the server <b>104</b> reads data from a storage area of a storage system <b>106</b> and transfers the data to another storage system <b>106</b>. In a data copy in a storage system <b>106</b>, the server <b>104</b> issues an instruction that data is read from a storage area of the storage system <b>106</b> and is stored the data in another area of the storage system <b>106</b>. In a data striping operation, the server <b>104</b> divides data stored in a storage system <b>106</b> into data items and sequentially transfers the data items to a plurality of disk devices of the storage system <b>106</b>. Moreover, the server generates redundant data according to the divided data items and stores the redundant data in the storage system. In a mirroring operation, the virtualization server <b>104</b> receives data from a server <b>101</b> and stores the data in a plurality of disk devices of the storage system <b>106</b> (<b>1504</b>).

P-0124[0124] When the processing is finished, the virtualization server <b>104</b> sends the processed data to the virtualization switch <b>105</b> (<b>1505</b>). Having transmitted the data, the server <b>104</b> returns to step <b>1502</b> to continue processing. If it is not determined in step <b>1503</b> that the received data is to be processed by the virtualization server <b>104</b>, the server <b>104</b> returns the data to the virtualization switch <b>105</b> (<b>1505</b>). The server <b>104</b> waits for reception of data in step <b>1502</b>.

P-0125[0125] If a plurality of virtualization switches are disposed, the virtualization server <b>104</b> selects according to the virtual volume control table <b>304</b> one of the switches <b>105</b> to transfer the data to the selected switch <b>105</b>.

P-0126[0126] In the description of the embodiment, the control server <b>102</b> is separated from the virtualization node <b>103</b>. However, the virtualization node <b>103</b> may include the control server <b>102</b>. Additionally, either one of a plurality of virtualization servers <b>104</b> included in the virtualization node <b>103</b> may operate as a control server <b>102</b>.

P-0127[0127] According to the embodiment, the virtualization node <b>103</b> to execute the virtualization processing includes at least one virtualization switch <b>105</b> to execute a high-speed processing in which, for example, the switch <b>105</b> receives data and converts the data into data of a real volume to transfer the data to an associated storage system <b>106</b> and the switch <b>105</b> received data from a storage system <b>106</b> and converts the data into data from a virtualized volume to send the data to a server <b>101</b> and at least one virtualization server <b>104</b> which executes complex processing in which, for example, the server <b>104</b> copies data between the storage systems <b>106</b>. Therefore, the virtualization processing in the virtualization node <b>103</b> can be executed in a sharing mode. As a result, there can be provided a virtualization system in which the virtualization node <b>103</b> does not become a bottleneck of the overall system processing.

P-0128[0128] According to the embodiment, the control server <b>102</b> which controls a correspondence between real volumes and virtual volumes and which controls the state of each device of the virtual system detects a virtualization switch <b>105</b> or a virtualization server <b>104</b> having a high processing load. The control server <b>102</b> conducts a control operation such that some virtual volumes being processed by the device in which a predetermined load has been detected are moved to a device having a low processing load. As a result, there can be configured a virtualization system in which the configuration of the virtualization system is changed according to increase in the processing load in the system to thereby provide stable performance.

P-0129[0129] According to the embodiment, the virtualization node <b>103</b> notifies increase in the processing load thereof to the control server <b>102</b>. In response to the notification, the control server <b>102</b> moves several volumes being processed by devices in the virtualization node <b>103</b> to other devices having a low processing load in the virtualization node <b>103</b>. As a result, it is possible to implement a virtualization system in which the virtualization is changed in the computer system according to increase in the processing load in the system to thereby provide performance at least equal to predetermined performance.

P-0130[0130] According to the embodiment, the control server <b>102</b> moves processing of a virtual volume being processed by a device in a failed virtualization node <b>103</b> to another device in a normal state. Therefore, there can be implemented a virtualization system in which even if a failure occurs in the virtualization system, a device in which an error occurs is easily replaced with a normal device in the virtualization system and hence the failure does not adversely influence any external device.

P-0131[0131] According to the embodiment, the virtualization node <b>103</b> notifies a failure thereof to the control server <b>102</b>. In response to the notification, the control server <b>102</b> moves processing of a virtual volume under processing to a device in a normal state. Therefore, there can be implemented a virtualization system in which even if an error occurs in the virtualization system, a failed device is easily replaced with a normal device in the virtualization system and hence the error does not adversely influence any external device.

P-0132[0132] It should be further understood by those skilled in the art that although the foregoing description has been made on embodiments of the invention, the invention is not limited thereto and various changes and modifications may be made without departing from the spirit of the invention and the scope of the appended claims.