CN1719785A - Method and system for monitoring performance of large scale memory system based on storage area network - Google Patents
Method and system for monitoring performance of large scale memory system based on storage area network Download PDFInfo
- Publication number
- CN1719785A CN1719785A CN 200510089929 CN200510089929A CN1719785A CN 1719785 A CN1719785 A CN 1719785A CN 200510089929 CN200510089929 CN 200510089929 CN 200510089929 A CN200510089929 A CN 200510089929A CN 1719785 A CN1719785 A CN 1719785A
- Authority
- CN
- China
- Prior art keywords
- storage
- performance
- area network
- storage area
- mss
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 85
- 238000012544 monitoring process Methods 0.000 title claims description 36
- 230000004044 response Effects 0.000 claims abstract description 8
- 238000004891 communication Methods 0.000 claims description 84
- 238000004364 calculation method Methods 0.000 claims description 28
- 238000013507 mapping Methods 0.000 claims description 5
- 230000003068 static effect Effects 0.000 claims description 4
- 102100027685 Hemoglobin subunit alpha Human genes 0.000 description 48
- 101001009007 Homo sapiens Hemoglobin subunit alpha Proteins 0.000 description 48
- 238000007726 management method Methods 0.000 description 32
- 238000005516 engineering process Methods 0.000 description 26
- 239000013307 optical fiber Substances 0.000 description 25
- 101100338341 Xenopus laevis hba3 gene Proteins 0.000 description 17
- 230000005540 biological transmission Effects 0.000 description 12
- 230000008859 change Effects 0.000 description 10
- 230000008901 benefit Effects 0.000 description 9
- 238000013480 data collection Methods 0.000 description 9
- 238000010586 diagram Methods 0.000 description 9
- 230000006855 networking Effects 0.000 description 7
- 238000011161 development Methods 0.000 description 6
- 230000035772 mutation Effects 0.000 description 6
- 239000000835 fiber Substances 0.000 description 5
- 238000005065 mining Methods 0.000 description 5
- 238000012856 packing Methods 0.000 description 5
- 230000003247 decreasing effect Effects 0.000 description 4
- 238000004422 calculation algorithm Methods 0.000 description 3
- 238000000151 deposition Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 230000009897 systematic effect Effects 0.000 description 3
- 238000012546 transfer Methods 0.000 description 3
- 230000015572 biosynthetic process Effects 0.000 description 2
- 241001269238 Data Species 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013523 data management Methods 0.000 description 1
- 230000001934 delay Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000013439 planning Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 108090000623 proteins and genes Proteins 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Images
Landscapes
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
This invention provides a performance monitor method for mass memory system based on storage domain network including setting up a parameter computing model containing response time, throughput, wasting time and service time for a mass memory system based on a storage domain network by comparison/image in a performance monitor, which utilizes the I/O required information of a storage server from the storage domain server as the input of a parameter computing model and computes the model to predict the performance target values of the original mass system and takes it as the monitor result to be output to manage the system. A monitor system is also provided including an information collector for collecting I/O required information of service ends, a performance monitor storing parameter computing models and a secondary network.
Description
Technical field
The invention belongs to performance of storage system monitoring field, be specifically related to a kind of method for monitoring performance and system of the mass storage system (MSS) based on storage area network.
Background technology
The promotion that the develop rapidly of ecommerce in recent years,, Digital Media and science data management (as fields such as the human gene of frontier nature, climatic analysises) is rapid to the demand of mass storage system (MSS).Storage area network is as a kind of open storage system architecture, the connection framework of tool switch type can be supported the flexible expansion of disposal ability, memory capacity and data access bandwidth, can satisfy the burgeoning unlimited demand (the true extension ability is relevant with the signal attenuation situation with the addressing support of equipment) that is stored in the extensibility aspect in theory.Therefore, in above-mentioned field, especially over past ten years, the architecture of storage system is the storage area network structure from traditional direct-connected structural transformation just gradually in recent years.Yet, although depend on the possibility that thereby advanced infrastructure has provides the high-performance stores service based on the mass storage system (MSS) of storage area network technology, but because the structure of storage area network relative complex and system scale for the conventional store mode is usually bigger, may cause systematic function to reduce operation even make mistakes, collapse if its performance is not carried out rational performance management, and then make the investment of system's construction waste, normal service can not be provided.Therefore how the mass storage system (MSS) based on the storage area network technology is carried out that rational performance management makes it stable and the work of high-performance ground is more and more paid close attention to by the insider, and cost-effective method for monitoring performance is the basis and the prerequisite of carrying out rational performance management.
At present, (also claim " storage QoS (service quality) monitoring and management " in document that has or the document for method for monitoring performance based on the mass storage system (MSS) of storage area network, in view of " storage QoS " also clearly not generally acknowledged at present definition, may be different in different local implications, for the purpose of clear and definite, we still claim " method for monitoring performance of storage system ") can be divided into two classes: outer (out-of-band) method of (in-band) method and band in the band.Method in the so-called band assigns into the performance data collection device exactly on the difference in I/O path and carries out performance monitoring.Collector its port between Data transmission bag or frame, and meanwhile carry out the statistics of performance data.By the statistics of these performance datas, can obtain some valuable performance index values, for example throughput, response time, delays etc. are to be used for performance management.So-called out-of-band method is exactly by direct visit storage area network equipment but not assign collector to come collecting performance data to obtain performance index value on the I/O path.In order to visit storage area network equipment, the out of band supervision device must use the exclusive API of equipment (application programming interfaces) to come the equipment different with each to carry out communication by time net (secondary network).
The advantage of method is to need not to visit API (because the opening of storage area network of each distinct device itself in the band, the equipment that constitutes storage area network can be from different manufacturers, and do not have general performance management interface at present between the different vendor, distinct device), the equipment in this method and the storage area network is independent of each other.The advantage of out-of-band method is can not bring extra delay, because this method needn't be assigned collector on the I/O path.
Shown in Figure 1 is that an interior method of employing band is to carrying out the example of performance monitoring based on the mass storage system (MSS) of storage area network.In this embodiment, the trunk of storage area network (heavy line among the figure) is made up of two optical fiber switch FCSW1 and FCSW2 interconnection, one of them optical fiber switch FCSW1 connects two storage server Serv1 and Serv2, and another optical fiber switch FCSW2 connects a storage server Serv3 and two optical fiber memory device S1 and S2.Client computer by communication subnet visit storage server Serv1, Serv2 or Serv3 with the stores service of obtaining optical fiber memory device S1 in the storage area network or S2 (, so in Fig. 1, omitting) because communication subnet, client computer are irrelevant with content of the present invention.Performance data collection device C1, C2, C3, C4, C5 and C6 are assigned into each section in I/O path, and packet (or Frame) advances from another port from the port of performance data collection device C1, C2, C3, C4, C5 and C6.The work of performance data collection device C1, C2, C3, C4, C5 and C6 is exactly to carry out the performance data statistics by the packet (or Frame) that transmits its place section.For other equipment in the storage area network, performance data collection device C1, C2, C3, C4, C5 and C6 are fully transparent, because performance data collection device C1, C2, C3, C4, C5 and C6 do not carry out any modification to the packet (or Frame) by it, only check (supervision) bag (or frame) head and afterbody for information about, for example source address, destination address, bag (or frame) size etc.In the method, Network Performance Monitor M uses unified API to come to carry out communication with each performance data collection device C1, C2, C3, C4, C5 and C6 by time net (fine dotted line among the figure adopts the IP network technology to realize usually) in band.Network Performance Monitor M calculates relevant performance index by association from the information of each collector in the band, as response time, throughput, time-delay etc.
The advantage of method is that it and monitored storage area network equipment are completely independent from one another and transparent in the band, need not to visit the API of each distinct device in the storage area network and need not to rely on these equipment and carry out performance data collection.Therefore be with interior method except the trouble of release, and be a kind of comparatively just method with the distinct device communication.But its disadvantage is each section that collector is assigned into the I/O path and can introduces extra transmission delay, the storage area network performance itself is brought influence.Although method is not carried out any retouching operation, only just checks and transmitted packet (or Frame) in the band, other equipment are less relatively in the delay that collector is introduced, and the delay that brings after all has certain influence to systematic function itself.
Shown in Figure 2 is that an employing out-of-band method is to carrying out the example of performance monitoring based on the mass storage system (MSS) of storage area network.Identical with the structure of storage area network among Fig. 1, in this embodiment, the trunk of storage area network (heavy line among the figure) is made up of two optical fiber switch FCSW1 and FCSW2 interconnection, one of them optical fiber switch FCSW1 connects two storage server Serv1 and Serv2, and another optical fiber switch FCSW2 connects a storage server Serv3 and two optical fiber memory device S1 and S2.Client computer by communication subnet visit storage server Serv1, Serv2 or Serv3 with the stores service of obtaining optical fiber memory device S1 in the storage area network or S2 (, so in Fig. 2, omitting) because communication subnet, client computer are irrelevant with content of the present invention.Out-of-band method does not have to adopt special collector to come collecting performance data as method in the band, but do as one likes can monitors M directly come the equipment different with each to carry out communication with collecting performance data by time net (fine dotted line among the figure adopts the IP network technology to realize usually) by the exclusive API of distinct device (comprising optical fiber switch FCSW1 and FCSW2, storage server Serv1, Serv2 and Serv3 and optical fiber memory device S1 and S2 in this example).Because the opening of storage area network itself, the equipment of formation storage area network can be from different manufacturers, and do not have general performance management API at present between the different vendor, distinct device.Be with outer Network Performance Monitor M to calculate relevant performance index value from the information of each equipment, as response time, throughput, time-delay etc. by association.
The advantage of out-of-band method is not have collector to get involved the I/O path, and the collection of performance data is finished by time net independently fully, therefore can not introduce extra transmission delay.But its disadvantage is each equipment that it need rely in the storage area network and carries out performance data collection, and is opaque to each equipment of monitored storage area network; Need to support and visit the relevant API of each distinct device in the storage area network simultaneously, when renewal of the equipment or upgrading, be with outer Network Performance Monitor also may need to upgrade accordingly especially.Therefore out-of-band method can't be exempted the trouble with the distinct device communication, compares independent and just inadequately with method in the band.
Although above-mentioned two kinds of methods each have the pluses and minuses of himself, and under each comfortable different situation the successfully case of utilization is arranged, they have one to limit the common drawback that it is accepted extensively and implements: implementation cost is high.For method in the band, need to buy, install, test and safeguard a large amount of collector (quantity of collector along with the complexity of storage area network topological structure sharply increase), and also need the right collector of different ports for the section that adopts different interconnection techniques in the I/O path (connecting, also may adopt High Speed IP to connect) as adopting optical fiber FC.For out-of-band method, need to support and safeguard with from the relevant API of different vendor, distinct device communication by time netting.And these two kinds of methods all need to guarantee constantly and safeguard time net safe operation, and the scale of time net will constantly increase along with the complexity of storage area network topological structure simultaneously.These above-mentioned work all need the implementer to employ great amount of manpower and financial resources.Even and if large corporation employs great amount of manpower and financial resources have been implemented one of these two kinds of methods, the performance monitoring system based on the mass storage system (MSS) of storage area network of the costliness of gained also is (fragile) that is easy to lose efficacy: the inefficacy that any point mistake in this numerous and jumbled system in related hardware, software and the manpower or error all can cause the performance monitoring result, and then may cause serious consequence (as: storage system fallback, stores service make mistakes and even collapse).
Summary of the invention
Above-mentioned shortcoming in view of prior art, the invention provides a kind of method of carrying out performance monitoring based on the mass storage system (MSS) of storage area network, it is characterized in that, this method comprises the steps: step 1 successively, in Network Performance Monitor, set up the calculation of parameter model that comprises response time, throughput, stand-by period, service time for mass storage system (MSS) based on storage area network by contrast/mapping means; Step 2, described Network Performance Monitor utilization are gathered the input of the I/O solicited message of storage server in storage area network as described calculation of parameter model; Step 3, described Network Performance Monitor calculates described calculation of parameter model, dope every performance index value of former mass storage system (MSS) based on storage area network, described performance index value is exported as the performance monitoring result, be used for the mass storage system (MSS) based on storage area network is carried out performance management, described mass storage system (MSS) based on storage area network is by storage server group, storage subnet, constitute with the storage subsystem that described storage server group links to each other by the storage subnet, and client computer links to each other with the storage server of correspondence in the described storage server group by communication subnet.
The invention is characterized in,
The signal of method for monitoring performance of the present invention such as Fig. 3.Storage server Serv1 among Fig. 3 ... Servm, storage subsystem S1 ... Sp and the mass storage system (MSS) SANsys (frame of broken lines in) of storage subnet StorNet formation based on storage area network, client computer C1 ... Cn is by communication subnet ComNet visit storage server Serv1 ... Servm is to obtain based on storage subsystem S1 among the mass storage system (MSS) SANsys of storage area network ... the stores service of Sp.The method of the invention is exactly at first the mass storage system (MSS) SANsys based on storage area network to be set up the calculation of parameter model by contrast/mapping means, utilize and gather from storage server Serv1 ... the I/O solicited message of Servm is as the input of above-mentioned parameter computation model, then described calculation of parameter model is calculated, dope every performance index value of former mass storage system (MSS) based on storage area network, described performance index value as performance monitoring result output, is used for the mass storage system (MSS) based on storage area network is carried out performance management.
The static parameter initial value of model is provided with according to the capability configuration parameters of real system in the said method.If adopt the automatic device discovery technology in the SRM instrument, then modeling and model initial parameter be provided with can realize automatic or semi-automatic.The I/O solicited message of the input of model---storage server can be gathered in the operating system layer or the application layer of storage server end.This collecting work need not the interior method of image-tape needs special-purpose collector to do hardware supports like that, only need take some cpu instruction times of implementation seldom on the server.Based on the performance performance of the mass storage system (MSS) of storage area network is to determine by system itself (comprising topological structure and system configuration) and as the I/O request operating load of system's input.The calculation of parameter model is to the annexation of system topology and system configuration and quantificational description, and the I/O solicited message of gathering is to being the attribute and the quantificational description of I/O request operating load.Therefore with the input of correct I/O solicited message as the reasonable parameter computation model, the performance index value that calculates will reasonably reflect the situation of systematic function.
According to an optimal way of the present invention, as the described I/O solicited message of described calculation of parameter model input be setting-up time at interval in to storage area network in the I/O of storage server ask the object information added up.
Adopt this method for monitoring performance based on the calculation of parameter model, because the calculating of analytical model all advantageous on time and space (emulation mode in time and the modeling of space complexity relative performance is lower), so the enforcement of this method need not to adopt the Network Performance Monitor of high configuration just can calculate every performance index value in real time.These characteristics provide higher flexibility for the enforcement of system: if performance manager is with the form appearance of private server, then Network Performance Monitor can be used as one of them application program, and the calculating of calculation of parameter model just can be finished by this application program; If Network Performance Monitor occurs with the form of specialized hardware facility, then Network Performance Monitor can be used as the firmware program in the embedded system, and the calculating of calculation of parameter model just can be finished by it.In addition, because the calculation of parameter model be to system at the logical abstraction of aspect of performance, and can regard a kind of Forecasting Methodology as based on statistics, therefore adopt some noise jamming that in preceding two kinds of methods, may occur of eliminating that this method can also " discard the dross and select the essential " (as follows net is temporary transient lost efficacy or acquired signal is accidental makes mistakes), i.e. stalwartness (robust) relatively.
The present invention also provides a kind of and carries out the system of performance monitoring based on what said method was set up based on the mass storage system (MSS) of storage area network, and it is characterized in that this system comprises: information acquisition device is used for acquisition server end I/O solicited message; Network Performance Monitor, having by contrast/mapping means in this watch-dog is the described calculation of parameter model of setting up based on the mass storage system (MSS) of storage area network that comprises response time, throughput, stand-by period, service time, the static parameter initial value of described calculation of parameter model is set according to the capability configuration parameters of described mass storage system (MSS), the information that above-mentioned information acquisition device is gathered is imported as described calculation of parameter model, and the output performance desired value is as monitored results; Inferior net, this time net are the communication networks that connects described information acquisition device and described Network Performance Monitor, are used for the information that described information acquisition device is gathered is delivered to described Network Performance Monitor; Described mass storage system (MSS) based on storage area network is by storage server group, storage subnet, constitute with the storage subsystem that described storage server group links to each other by the storage subnet, and client computer links to each other with the storage server of correspondence in the described storage server group by communication subnet.
Example such as Fig. 4 of adopting that this method makes up based on the performance monitoring system of the mass storage system (MSS) of storage area network.In this embodiment, the structure of storage area network is identical with the example of Fig. 1 and Fig. 2: the trunk of storage area network (heavy line among the figure) is made up of two optical fiber switch FCSW1 and FCSW2 interconnection, one of them optical fiber switch FCSW1 connects two storage server Serv1 and Serv2, and another optical fiber switch FCSW2 connects a storage server Serv3 and two optical fiber memory device S1 and S2.Client computer by communication subnet visit storage server Serv1, Serv2 or Serv3 with the stores service of obtaining optical fiber memory device S1 in the storage area network or S2 (, so in Fig. 4, omitting) because communication subnet, client computer are irrelevant with content of the present invention.In the method, Network Performance Monitor M is by time net (fine dotted line among the figure, usually adopt the IP network technology to realize) only need to gather the I/O solicited message from storage server Serv1, Serv2 and Serv3, because of the scale of this time net all little more a lot of than method and out-of-band method in the aforesaid band: with the out-of-band method ratio, inferior net need not to connect exchanging interconnection equipment (being optical fiber switch FCSW1 and FCSW2 in this example) and the storage subsystem equipment (being optical fiber memory device S1 and S2 in this example) in the storage subnet; With method ratio in the band, the storage server number that inferior net connected only equates to the I/O path hop count of storage between the subnet with storage server, thus can save the storage subnet inner with store being connected of the collector of subnet to the I/O route segment of storage subsystem equipment.Simultaneously, this method only needs to gather the I/O solicited message from storage server Serv1, Serv2 and Serv3, and this information can obtain by the association request message of intercept operation system layer or application layer.And for mainstream operation system and application layer framework, the interface relevant with request message is compatible or similar.
Therefore, Network Performance Monitor need not Network Performance Monitor in the outer method of image-tape and supports different API from various distinct devices like that in this method.And, this method is only visited each storage server and is gathered the I/O solicited message, gathering action does not relate to exchanging interconnection equipment in the storage area network and does not relate to storage subsystem yet, therefore also littler to the performance impact of system itself than out-of-band method, the farther time delay that causes much smaller than the collector of being with hardware form in the interior method.At last, because this method does not rely on the exchanging interconnection equipment in the storage area network and the relevant API of storage subsystem equipment, its independence and fairness are better than out-of-band method.
To sum up, major advantage of the present invention is: 1) significantly reduce time network planning mould, 2) less to the performance impact of system own, 3) need not to support different API from various distinct devices, 4) independence and fairness are better than out-of-band method, 5) lower, the environment that is suitable for having relatively high expectations in real time of time and space complexity, 6) healthy and strong relatively.Therefore, compare with out-of-band method with method in the aforementioned band, this method is not only simple in structure, comparatively healthy and strong (robust), and implementation cost is also cheap relatively.
Description of drawings
Fig. 1 is a method schematic diagram in a kind of band of the prior art;
Fig. 2 is a kind of out-of-band method schematic diagram of the prior art;
Fig. 3 is a method for monitoring performance schematic diagram of the present invention;
Fig. 4 is the schematic diagram according to the constructed performance monitoring system of the inventive method;
Fig. 5 is the schematic diagram according to the performance monitoring system of first embodiment of the invention;
Fig. 6 is the schematic diagram according to the performance monitoring system of second embodiment of the invention;
Fig. 7 is the schematic diagram according to the performance monitoring system of third embodiment of the invention;
Fig. 8 is the schematic diagram according to the performance monitoring system of fourth embodiment of the invention;
Fig. 9 is the schematic diagram according to the performance monitoring system of fifth embodiment of the invention.
Embodiment
Describe each embodiment of the present invention in detail below in conjunction with accompanying drawing.
Performance monitoring system according to the inventive method mainly comprises 3 parts: server end I/O solicited message collector (acquisition module); Network Performance Monitor is used to handle the information that collector is gathered, and the method for processing is to do operation rule and environment, the information of gathering is done the computing input, calculated performance parameter value according to calculation of parameter model of setting and parameter thereof; Inferior net, its feature difference is the communication network that is used for the information of acquisition module collection is delivered to Network Performance Monitor in communication subnet and storage subnet.
In different embodiment, 3 chief components of the present invention (or form) are in a different manner realized.Listed the execution mode of 5 quasi-representatives below, the simple variant of this 5 class execution mode will illustrate when introducing separately embodiment.
First embodiment is described below, and wherein, inferior net is that Network Performance Monitor is with the specialized hardware facility with independently network form appearance physically, such as the form appearance of single-chip microcomputer or special chip.
As shown in Figure 5, I/O solicited message collector (acquisition module) 1.1 ... each provides the server S erv.1 of storage related service App in the storage area network by communication subnet 1.n be installed in ... on the Serv.n.Each server S erv.1 ... 3 host bus adaptor are installed on the Serv.n, HBA1.1 ... HBA1.n, HBA2.1 ... HBA2.n and HBA3.1 ... HBA3.n is respectively applied for server S erv.1 ... Serv.n with physically independently store being connected and communication (having omitted communication subnet and client computer among Fig. 5) of subnet 4, inferior net 3 and communication subnet respectively with the no substantial connection of present embodiment explanation.Each storage subsystem S.1 ... 1 host bus adaptor HBA1.s.1 is installed on the S.m ... S.1 HBA1.s.m is used for storage subsystem ... S.m is connected and communication with storage subnet 4.Host bus adaptor HBA1.1 ... HBA1.n and HBA1.s.1 ... HBA1.s.m is according to the interconnection technique scheme difference of storage subnet 4, can be that the fibre channel host bus adapter is (corresponding to the optical-fibre channel storage area network, be FC SAN), High Speed IP host's machine bus adapter (corresponding to the IP storage area network, i.e. IP SAN) or other host bus adaptor.
Because transmission only is the property value of I/O solicited message on the inferior net 3, the institute in proper order common IP networking technology of net 3 employings can satisfy the demand of transmission bandwidth.Host bus adaptor HBA2.0, HBA2.1 ... HBA2.n according to the IP technical schemes difference of inferior net 3, can be wired ethernet card (corresponding to wired IP network) or wireless ethernet card (corresponding to Wireless IP) as the part of time net 3.Host bus adaptor HBA3.1 ... HBA3.n uses the host bus adaptor of relevant art according to the networking technology different mining of communication subnet.I/O solicited message collector (acquisition module) 1.1 ... 1.n be responsible for monitoring service device Serv.1 ... operating system OS is to host bus adaptor HBA1.1 among the Serv.n ... the I/O request that HBA1.n sends, and with the packing of the association attributes of I/O solicited message, by inlet (the host bus adaptor HBA2.1 of inferior net 3 ... HBA2.n) be sent to Network Performance Monitor 2.Inferior net 3 comprises three parts: be positioned at server S erv.1 ... the host bus adaptor HBA2.1 of Serv.n ... HBA2.n, be arranged in Network Performance Monitor 2 host bus adaptor HBA2.0, connect all time net 3 host bus adaptor HBA2.0, HBA2.1 ... the exchanging interconnection equipment of HBA2.n and connection cable (wired network) or wireless channel (wireless network).Network Performance Monitor 2 is connected to time net 3 by host bus adaptor HBA2.0, and accepts I/O solicited message collector (module) 1.1 by this host bus adaptor HBA2.0 ... 1.n the association attributes bag of the I/O solicited message of sending.
In the present embodiment, Network Performance Monitor 2 occurs with the form (single-chip microcomputer) of specialized hardware facility, it is except comprising the aforementioned host bus adaptor HBA2.0 and the channel interface 24 that is used for to performance manager transmission performance indicators value that is used for time net 3 communications, also comprise 3 parts that link to each other by system bus: CPU 21, RAM 22 and EPROM23 provide the storage of computing capability, computer memory, performance model characterising parameter and performance model computations respectively.The renewal of performance model characterising parameter and performance model computations realizes by refreshing EPROM 23, and this change that makes calculation of parameter model and parameter thereof cause along with the change of storage area network equipment or improving of modeling technique can be included present embodiment easily and flexibly in.Adopt general-purpose computations machine board card, interface and adapter in the present embodiment, help absorbing IT technology new results, make the mutation of such embodiment can reach the purpose that can constantly reduce cost and improve performance quality along with the IT development of technology in future.
Second embodiment is described below, and wherein, inferior net is with independently network form appearance physically, and Network Performance Monitor is the form appearance with application program in the special-purpose performance management server.
As shown in Figure 6, I/O solicited message collector (acquisition module) 1.1 ... each provides the server S erv.1 of storage related service App in the storage area network by communication subnet 1.n be installed in ... on the Serv.n.Each server S erv.1 ... 3 host bus adaptor are installed on the Serv.n, HBA1.1 ... HBA1.n, HBA2.1 ... HBA2.n and HBA3.1 ... HBA3.n is respectively applied for server S erv.1 ... Serv.n with physically independently store being connected and communication (having omitted communication subnet and client computer among Fig. 6) of subnet 4, inferior net 3 and communication subnet respectively with the no substantial connection of present embodiment explanation.Each storage subsystem S.1 ... 1 host bus adaptor HBA1.s.1 is installed on the S.m ... S.1 HBA1.s.m is used for storage subsystem ... S.m is connected and communication with storage subnet 4.Host bus adaptor HBA1.1 ... HBA1.n and HBA1.s.1 ... HBA1.s.m is according to the interconnection technique scheme difference of storage subnet 4, can be that the fibre channel host bus adapter is (corresponding to the optical-fibre channel storage area network, be FC SAN), High Speed IP host's machine bus adapter (corresponding to the IP storage area network, i.e. IP SAN) or other host bus adaptor.
Because transmission only is the property value of I/O solicited message on the inferior net 3, the institute in proper order common IP networking technology of net 3 employings can satisfy the demand of transmission bandwidth.Host bus adaptor HBA2.0, HBA2.1 ... HBA2.n according to the IP technical schemes difference of inferior net 3, can be wired ethernet card (corresponding to wired IP network) or wireless ethernet card (corresponding to Wireless IP) as the part of time net 3.Host bus adaptor HBA3.1 ... HBA3.n uses the host bus adaptor of relevant art according to the networking technology different mining of communication subnet.I/O solicited message collector (acquisition module) 1.1 ... 1.n be responsible for monitoring service device Serv.1 ... operating system OS is to HBA1.1 among the Serv.n ... the I/O request that HBA1.n sends, and with the packing of the association attributes of I/O solicited message, the inlet by inferior net 3 (is host bus adaptor HBA2.1 ... HBA2.n) be sent to special-purpose performance management server 5.Inferior net 3 comprises three parts: be positioned at server S erv.1 ... the host bus adaptor HBA2.1 of Serv.n ... HBA2.n, be arranged in special-purpose performance management server 5 host bus adaptor HBA2.0, connect all time net 3 host bus adaptor HBA2.0, HBA2.1 ... the exchanging interconnection equipment of HBA2.n and connection cable (wired network) or wireless channel (wireless network).Special-purpose performance management server 5 is connected to time net 3 by host bus adaptor HBA2.0, and accepts I/O solicited message collector (module) 1.1 by this host bus adaptor HBA2.0 ... 1.n the association attributes bag of the I/O solicited message of sending.
In the present embodiment, Network Performance Monitor 52 occurs with the application program form that resides in the special-purpose performance management server 5, the performance index value that calculates is directly passed to the performance management application that resides in equally in the special-purpose performance management server 51 and uses, and transfer mode can adopt: shared data structure, shared file, shared data bank, application programming interfaces call, message transmission etc. is multiple with communication modes between application program in the machine.In the present embodiment, performance model and parametric description thereof are stored in its corresponding description document (or table) 53.The calculating do as one likes of performance model can supervisor applications program 52 be finished.The renewal of performance model and characterising parameter thereof and performance model computational algorithm is to realize with the correlative code of Network Performance Monitor application program 52 by upgrading performance model and parametric description file 53, and this change that makes calculation of parameter model and parameter thereof cause along with the change of depositing territory equipment or improving of modeling technique can be included present embodiment easily and flexibly in.Adopt general-purpose computations machine board card, interface and adapter in the present embodiment, help absorbing IT technology new results, make the mutation of such embodiment can reach the purpose that can constantly reduce cost and improve performance quality along with the IT development of technology in future.
The 3rd embodiment is described below, wherein, inferior net be use communication subnet, with physically not independently the network form occur, Network Performance Monitor is that the form with application program in the special-purpose performance management server occurs.
As shown in Figure 7, I/O solicited message collector (acquisition module) 1.1 ... each provides the server S erv.1 of storage related service App in the storage area network by communication subnet 1.n be installed in ... on the Serv.n.Each server S erv.1 ... 2 host bus adaptor are installed on the Serv.n, HBA1.1 ... HBA1.n and HBA3.1 ... HBA3.n is respectively applied for server S erv.1 ... Serv.n with physically independently store subnet 4, communication subnet hold concurrently being connected of time net 6 and communication (having omitted client computer among Fig. 7) respectively with the no substantial connection of present embodiment explanation.Each storage subsystem S.1 ... 1 host bus adaptor HBA1.s.1 is installed on the S.m ... S.1 HBA1.s.m is used for storage subsystem ... S.m is connected and communication with storage subnet 4.Host bus adaptor HBA1.1 ... HBA1.n and HBA1.s.1 ... HBA1.s.m is according to the interconnection technique scheme difference of storage subnet 4, can be that the fibre channel host bus adapter is (corresponding to the optical-fibre channel storage area network, be FC SAN), High Speed IP host's machine bus adapter (corresponding to the IP storage area network, i.e. IP SAN) or other host bus adaptor.Host bus adaptor HBA3.0, HBA3.1 ... HBA3.n for example can be wired ethernet card (corresponding to wired IP network) or wireless ethernet card (corresponding to Wireless IP) according to the host bus adaptor of the networking technology different mining usefulness relevant art of communication subnet.
In the present embodiment, the double time net of doing of communication subnet, its advantage is to save in a large number time implementation cost of net, its shortcoming is that part communication subnet bandwidth is occupied, performance to communication subnet is influential, and the network congestion of communication subnet may cause time net decreased performance even packet loss conversely.This embodiment is applicable to the situation that communication subnet is not busy, bandwidth resources are idle, does not recommend to adopt for the busy situation of communication subnet.
I/O solicited message collector (acquisition module) 1.1 ... 1.n be responsible for monitoring service device Serv.1 ... operating system OS is to host bus adaptor HBA1.1 among the Serv.n ... the I/O request that HBA1.n sends, and with the packing of the association attributes of I/O solicited message, by communication subnet inlet (the host bus adaptor HBA3.1 of time net 6 that holds concurrently ... HBA3.n) be sent to special-purpose performance management server 5.Special-purpose performance management server 5 is connected to the double time net 6 of communication subnet by host bus adaptor HBA3.0, and accepts I/O solicited message collector (module) 1.1 by this host bus adaptor HBA3.0 ... 1.n the association attributes bag of the I/O solicited message of sending.
In the present embodiment, Network Performance Monitor 52 occurs with the application program form that resides in the special-purpose performance management server 5, the performance index value that calculates is directly passed to the performance management application that resides in equally in the special-purpose performance management server 51 and uses, and transfer mode can adopt: shared data structure, shared file, shared data bank, application programming interfaces call, message transmission etc. is multiple with communication modes between application program in the machine.In the present embodiment, performance model and parametric description thereof are stored in its corresponding description document (or table) 53.The calculating do as one likes of performance model can supervisor applications program 52 be finished.The renewal of performance model and characterising parameter thereof and performance model computational algorithm is to realize with the correlative code of Network Performance Monitor application program 52 by upgrading performance model and parametric description file 53, and this change that makes calculation of parameter model and parameter thereof cause along with the change of depositing territory equipment or improving of modeling technique can be included present embodiment easily and flexibly in.Adopt general-purpose computations machine board card, interface and adapter in the present embodiment, help absorbing IT technology new results, make the mutation of such embodiment can reach the purpose that can constantly reduce cost and improve performance quality along with the IT development of technology in future.
The 4th embodiment is described below, wherein, inferior net be use communication subnet, with physically not independently the network form occur, Network Performance Monitor is with the specialized hardware facility, occurs such as the form of single-chip microcomputer or special chip.
As shown in Figure 8, I/O solicited message collector (acquisition module) 1.1 ... each provides the server S erv.1 of storage related service App in the storage area network by communication subnet 1.n be installed in ... on the Serv.n.Each server S erv.1 ... 2 host bus adaptor are installed on the Serv.n, HBA1.1 ... HBA1.n and HBA3.1 ... HBA3.n is respectively applied for server S erv.1 ... Serv.n with physically independently store subnet 4, communication subnet hold concurrently being connected of time net 6 and communication (having omitted client computer among Fig. 8) respectively with the no substantial connection of present embodiment explanation.Each storage subsystem S.1 ... 1 host bus adaptor HBA1.s.1 is installed on the S.m ... S.1 HBA1.s.m is used for storage subsystem ... S.m is connected and communication with storage subnet 4.Host bus adaptor HBA1.1 ... HBA1.n and HBA1.s.1 ... HBA1.s.m is according to the interconnection technique scheme difference of storage subnet 4, can be that the fibre channel host bus adapter is (corresponding to the optical-fibre channel storage area network, be FC SAN), High Speed IP host's machine bus adapter (corresponding to the IP storage area network, i.e. IP SAN) or other host bus adaptor.Host bus adaptor HBA3.0, HBA3.1 ... HBA3.n for example can be wired ethernet card (corresponding to wired IP network) or wireless ethernet card (corresponding to Wireless IP) according to the host bus adaptor of the networking technology different mining usefulness relevant art of communication subnet.
In the present embodiment, the double time net of doing of communication subnet, its advantage is to save in a large number time implementation cost of net, its shortcoming is that part communication subnet bandwidth is occupied, performance to communication subnet is influential, and the network congestion of communication subnet may cause time net decreased performance even packet loss conversely.This embodiment is applicable to the situation that communication subnet is not busy, bandwidth resources are idle, does not recommend to adopt for the busy situation of communication subnet.
I/O solicited message collector (acquisition module) 1.1 ... 1.n be responsible for monitoring service device Serv.1 ... operating system OS is to host bus adaptor HBA1.1 among the Serv.n ... the I/O request that HBA1.n sends, and with the packing of the association attributes of I/O solicited message, by communication subnet inlet (the host bus adaptor HBA3.1 of time net 6 that holds concurrently ... HBA3.n) be sent to Network Performance Monitor 2.Network Performance Monitor 2 is connected to the double time net 6 of communication subnet by host bus adaptor HBA3.0, and accepts I/O solicited message collector (module) 1.1 by this host bus adaptor HBA3.0 ... 1.n the association attributes bag of the I/O solicited message of sending.
In the present embodiment, Network Performance Monitor 2 occurs with the form (single-chip microcomputer) of specialized hardware facility, it is except comprising the aforementioned host bus adaptor HBA3.0 and the channel interface 24 that is used for to performance manager transmission performance indicators value that is used for double time net 6 communications of communication subnet, also comprise 3 parts that link to each other by system bus: CPU 21, RAM 22 and EPROM 23 provide the storage of computing capability, computer memory, performance model characterising parameter and performance model computations respectively.The renewal of performance model characterising parameter and performance model computations realizes by refreshing EPROM 23, and this change that makes calculation of parameter model and parameter thereof cause along with the change of storage area network equipment or improving of modeling technique can be included present embodiment easily and flexibly in.Adopt general-purpose computations machine board card, interface and adapter in the present embodiment, help absorbing IT technology new results, make the mutation of such embodiment can reach the purpose that can constantly reduce cost and improve performance quality along with the IT development of technology in future.
The 5th embodiment is described below, wherein, inferior net be use communication subnet, with physically not independently the network form occur, Network Performance Monitor is to use one of server, occur with the form of application program in the non-special-purpose performance management server.
As shown in Figure 9, I/O solicited message collector (acquisition module) 1.1 ... 1.i ... each provides the server S erv.1 of storage related service App in the storage area network by communication subnet 1.n be installed in ... Serv.i ... on the Serv.n.Each server S erv.1 ... Serv.i ... 2 host bus adaptor are installed on the Serv.n, HBA1.1 ... HBA1.i ... HBA1.n and HBA3.1 ... HBA3.i ... HBA3.n is respectively applied for server S erv.1 ... Serv.i ... Serv.n with physically independently store subnet 4, communication subnet hold concurrently being connected of time net 6 and communication (having omitted client computer among Fig. 9) respectively with the no substantial connection of present embodiment explanation.Each storage subsystem S.1 ... 1 host bus adaptor HBA1.s.1 is installed on the S.m ... S.1 HBA1.s.m is used for storage subsystem ... S.m is connected and communication with storage subnet 4.Host bus adaptor HBA1.1 ... HBA1.i ... HBA1.n and HBA1.s.1 ... HBA1.s.m is according to the interconnection technique scheme difference of storage subnet 4, can be that the fibre channel host bus adapter is (corresponding to the optical-fibre channel storage area network, be FCSAN), High Speed IP host's machine bus adapter (corresponding to the IP storage area network, i.e. IP SAN) or other host bus adaptor.Host bus adaptor HBA3.0, HBA3.1 ... HBA3.i ... HBA3.n for example can be wired ethernet card (corresponding to wired IP network) or wireless ethernet card (corresponding to Wireless IP) according to the host bus adaptor of the networking technology different mining usefulness relevant art of communication subnet.
In the present embodiment, the double time net of doing of communication subnet, its advantage is to save in a large number time implementation cost of net, its shortcoming is that part communication subnet bandwidth is occupied, performance to communication subnet is influential, and the network congestion of communication subnet may cause time net decreased performance even packet loss conversely.
I/O solicited message collector (acquisition module) 1.1 ... 1.i ... 1.n be responsible for monitoring service device Serv.1 ... Serv.i ... operating system OS is to host bus adaptor HBA1.1 among the Serv.n ... HBA1.i ... the I/O request that HBA1.n sends, and with the packing of the association attributes of I/O solicited message, by communication subnet inlet (the host bus adaptor HBA3.1 of time net 6 that holds concurrently ... HBA3.i ... HBA3.n) be sent to the server S erv.i of double as performance management server.The server S erv.i of double as performance management server is connected to the double time net 6 of communication subnet by host bus adaptor HBA3.i, and accepts I/O solicited message collector (module) 1.1 by this host bus adaptor HBA3.0 ... 1.i ... 1.n the association attributes bag of the I/O solicited message of sending.
In the present embodiment, Network Performance Monitor 52 occurs to reside in by the application program form among the double performance management server S erv.i that does of a server in the storage area network, the performance management application 51 that the performance index value that calculates is directly passed among this performance management server S erv.i is used, and transfer mode can adopt: shared data structure, shared file, shared data bank, application programming interfaces call, message transmission etc. is multiple with communication modes between application program in the machine.
In the present embodiment, by a double performance management server of doing of storage area network server, its advantage is to save the implementation cost of performance management server, its shortcoming is that the local calculating of the part of this storage area network server, storage and ability to communicate are occupied, performance to storage area network server itself is influential, and storage area network server self performance descends and may cause whole performance management system decreased performance conversely even make mistakes.This embodiment is applicable to 1) communication subnet is not busy, bandwidth resources idle, and 2) calculating of storage area network server, storage and the communication of double as performance management server have situation fully more than needed, for communication subnet busy or (with) situation of storage area network server resource limited (or busy) do not recommend to adopt.
In the present embodiment, performance model and parametric description thereof are stored in its corresponding description document (or table) 53.The calculating do as one likes of performance model can supervisor applications program 52 be finished.The renewal of performance model and characterising parameter thereof and performance model computational algorithm is to realize with the correlative code of Network Performance Monitor application program 52 by upgrading performance model and parametric description file 53, and this change that makes calculation of parameter model and parameter thereof cause along with the change of depositing territory equipment or improving of modeling technique can be included present embodiment easily and flexibly in.Adopt general-purpose computations machine board card, interface and adapter in the present embodiment, help absorbing IT technology new results, make the mutation of such embodiment can reach the purpose that can constantly reduce cost and improve performance quality along with the IT development of technology in future.
All adopt general-purpose computations machine board card, interface and adapter among the above-mentioned 5 class embodiment, help absorbing IT technology new results, make the mutation of above-mentioned 5 class embodiment can reach the purpose that can constantly reduce cost and improve performance quality along with the IT development of technology in future.
According to the explanation of above-mentioned 5 class execution modes, can select suitable execution mode for use according to concrete system scale, busy extent, economic budget.Summary is got up, and implementation cost is to successively decrease successively according to mode 1,2＞ mode 3,4＞mode 5, is to increase progressively successively according to mode 1,2＜ mode 3,4＜mode 5 to the influence of monitored system; And mode 1 and the main difference of mode 2 (mode 3 and mode 4) are the realization form of watch-dog, and the two should and require according to concrete enforcement environment choose at implementation cost, little to difference aspect the performance impact of system own.
Claims (12)
1. the method for monitoring performance based on the mass storage system (MSS) of storage area network is characterized in that, this method comprises the steps: successively
Step 1 in Network Performance Monitor, is set up the calculation of parameter model that comprises response time, throughput, stand-by period, service time by contrast/mapping means for the mass storage system (MSS) based on storage area network;
Step 2, described Network Performance Monitor utilization are gathered the input of the I/O solicited message of storage server in storage area network as described calculation of parameter model;
Step 3, described Network Performance Monitor calculates described calculation of parameter model, dope every performance index value of former mass storage system (MSS) based on storage area network, described performance index value is exported as the performance monitoring result, be used for the mass storage system (MSS) based on storage area network is carried out performance management, described mass storage system (MSS) based on storage area network is by storage server group, storage subnet, constitute with the storage subsystem that described storage server group links to each other by the storage subnet, and client computer links to each other with the storage server of correspondence in the described storage server group by communication subnet.
2. method according to claim 1 is characterized in that, the static parameter initial value of described calculation of parameter model is set according to the capability configuration parameters of monitored described mass storage system (MSS) based on storage area network.
3. method according to claim 1, it is characterized in that, as the described I/O solicited message of the input of described calculation of parameter model be setting-up time at interval in to described based on the mass storage system (MSS) of storage area network in the I/O of storage server ask the object information added up.
4. method according to claim 1 is characterized in that, described IO solicited message based on storage server in the mass storage system (MSS) of storage area network is gathered operating system layer or the application layer from described storage server.
5. performance monitoring system of setting up based on the described method of claim 1 based on the mass storage system (MSS) of storage area network is characterized in that this system comprises:
Information acquisition device is used for acquisition server end I/O solicited message;
Network Performance Monitor, having by contrast/mapping means in this watch-dog is the described calculation of parameter model of setting up based on the mass storage system (MSS) of storage area network that comprises response time, throughput, stand-by period, service time, the static parameter initial value of described calculation of parameter model is set according to the capability configuration parameters of described mass storage system (MSS), the information that above-mentioned information acquisition device is gathered is imported as described calculation of parameter model, and the output performance desired value is as monitored results;
Inferior net, this time net are the communication networks that connects described information acquisition device and described Network Performance Monitor, are used for the information that described information acquisition device is gathered is delivered to described Network Performance Monitor; Described mass storage system (MSS) based on storage area network is by storage server group, storage subnet, constitute with the storage subsystem that described storage server group links to each other by the storage subnet, and client computer links to each other with the storage server of correspondence in the described storage server group by communication subnet.
6. system according to claim 5 is characterized in that described information acquisition device is positioned at the operating system layer of the storage server of described mass storage system (MSS).
7. system according to claim 5 is characterized in that described information acquisition device is positioned at the application service layer of the storage server of described mass storage system (MSS).
8. system according to claim 5 is characterized in that, described Network Performance Monitor is the performance monitoring program in the special-purpose performance management server.
9. system according to claim 5 is characterized in that, described Network Performance Monitor is the performance monitoring program of using in the non-special-purpose performance management server of storage server of described mass storage system (MSS).
10. system according to claim 5 is characterized in that, described Network Performance Monitor is the specialized hardware facility.
11. system according to claim 5 is characterized in that, described net is the dedicated network that physically is independent of communication subnet and storage subnet.
12. system according to claim 5 is characterized in that, described net be use communication subnet, independently not non-dedicated network physically.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2005100899297A CN100347990C (en) | 2005-08-08 | 2005-08-08 | Method and system for monitoring performance of large scale memory system based on storage area network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2005100899297A CN100347990C (en) | 2005-08-08 | 2005-08-08 | Method and system for monitoring performance of large scale memory system based on storage area network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1719785A true CN1719785A (en) | 2006-01-11 |
| CN100347990C CN100347990C (en) | 2007-11-07 |
Family
ID=35931515
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2005100899297A Active CN100347990C (en) | 2005-08-08 | 2005-08-08 | Method and system for monitoring performance of large scale memory system based on storage area network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100347990C (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100428170C (en) * | 2007-01-26 | 2008-10-22 | 清华大学 | Load-aware IO performance optimization methods based on Bayesian decision |
| CN102638378A (en) * | 2012-02-22 | 2012-08-15 | 中国人民解放军国防科学技术大学 | Mass storage system monitoring method integrating heterogeneous storage devices |
| GB2514601A (en) * | 2013-05-30 | 2014-12-03 | Xyratex Tech Ltd | Method of, and apparatus for, detection of degradation on a storage resource |
| US9239746B2 (en) | 2013-05-30 | 2016-01-19 | Xyratex Technology Limited—A Seagate Company | Method of, and apparatus for, detection of degradation on a storage resource |
| CN111090401A (en) * | 2020-03-23 | 2020-05-01 | 上海飞旗网络技术股份有限公司 | Storage device performance prediction method and device |
| CN111756575A (en) * | 2020-06-19 | 2020-10-09 | 星辰天合（北京）数据科技有限公司 | Performance analysis method and device of storage server and electronic equipment |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6845344B1 (en) * | 2000-08-18 | 2005-01-18 | Emc Corporation | Graphical user input interface for testing performance of a mass storage system |
| CN1228726C (en) * | 2003-08-08 | 2005-11-23 | 华中科技大学 | Massive memory system based on multi-channel memory equipment and its control method |
-
2005
- 2005-08-08 CN CNB2005100899297A patent/CN100347990C/en active Active
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100428170C (en) * | 2007-01-26 | 2008-10-22 | 清华大学 | Load-aware IO performance optimization methods based on Bayesian decision |
| CN102638378A (en) * | 2012-02-22 | 2012-08-15 | 中国人民解放军国防科学技术大学 | Mass storage system monitoring method integrating heterogeneous storage devices |
| CN102638378B (en) * | 2012-02-22 | 2014-05-21 | 中国人民解放军国防科学技术大学 | Mass storage system monitoring method integrating heterogeneous storage devices |
| GB2514601A (en) * | 2013-05-30 | 2014-12-03 | Xyratex Tech Ltd | Method of, and apparatus for, detection of degradation on a storage resource |
| GB2514601B (en) * | 2013-05-30 | 2015-10-21 | Xyratex Tech Ltd | Method of, and apparatus for, detection of degradation on a storage resource |
| US9239746B2 (en) | 2013-05-30 | 2016-01-19 | Xyratex Technology Limited—A Seagate Company | Method of, and apparatus for, detection of degradation on a storage resource |
| CN111090401A (en) * | 2020-03-23 | 2020-05-01 | 上海飞旗网络技术股份有限公司 | Storage device performance prediction method and device |
| CN111090401B (en) * | 2020-03-23 | 2020-06-26 | 上海飞旗网络技术股份有限公司 | Storage device performance prediction method and device |
| CN111756575A (en) * | 2020-06-19 | 2020-10-09 | 星辰天合（北京）数据科技有限公司 | Performance analysis method and device of storage server and electronic equipment |
| CN111756575B (en) * | 2020-06-19 | 2023-08-11 | 北京星辰天合科技股份有限公司 | Performance analysis method and device of storage server and electronic equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN100347990C (en) | 2007-11-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Jeannot et al. | Near-optimal placement of MPI processes on hierarchical NUMA architectures | |
| US7440415B2 (en) | Virtual network addresses | |
| CN1212574C (en) | End node partitioning using local identifiers | |
| CN100347990C (en) | Method and system for monitoring performance of large scale memory system based on storage area network | |
| CN103392314B (en) | For the system and method that the N nuclear statistics information that can expand is polymerized | |
| MXPA04003956A (en) | Aggregate system resource analysis including correlation matrix and metric-based analysis. | |
| US8533524B2 (en) | Real time test result promulgation from network component test device | |
| CN1875585A (en) | Dynamic unknown L2 flooding control with MAC limits | |
| WO2012024909A1 (en) | Long connection management apparatus and link resource management method for long connection communication | |
| CN1838011A (en) | Intelligent management apparatus and management method for distributed control network based on CAN bus | |
| CN108400993A (en) | The Internet of things system and storage medium that intelligent industrial apparatus components formula is set up | |
| CN1794170A (en) | Tele communication region modeling tool based on unified modeling language and modeling method | |
| CN108833267A (en) | Intelligent industrial gateway | |
| CN105991361A (en) | Monitoring method and monitoring system for cloud servers in cloud computing platform | |
| CN115913937B (en) | Container multi-network card network configuration method, device, equipment and storage medium | |
| CN103684851B (en) | Collecting method and device | |
| CN116418700A (en) | Distributed data capturing method based on DPDK | |
| Iannaccone | Fast prototyping of network data mining applications | |
| Albrecht et al. | Making work queue cluster-friendly for data intensive scientific applications | |
| CN1917436A (en) | Method for realizing concentrating type management for network devices based on Web | |
| US20240070123A1 (en) | Using Machine Learning to Provide a Single User Interface for Streamlined Deployment and Management of Multiple Types of Databases | |
| Balman et al. | Experiences with 100gbps network applications | |
| CN1160906C (en) | Analogue web element testing method based on message recording and playback | |
| CN1780237A (en) | Logic topological network for protocol test and method for realizing protocol test | |
| CN1946035A (en) | Method for managing net element data configuration and net element |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| ASS | Succession or assignment of patent right |
Owner name: EASYWAY TECHNOLOGY CO., LTD. Free format text: FORMER OWNER: TSINGHUA UNIVERSITY Effective date: 20121225 |
|
| C41 | Transfer of patent application or patent right or utility model | ||
| COR | Change of bibliographic data |
Free format text: CORRECT: ADDRESS; FROM: 100084 HAIDIAN, BEIJING TO: 100083 HAIDIAN, BEIJING |
|
| TR01 | Transfer of patent right |
Effective date of registration: 20121225 Address after: 100083 Beijing city Haidian District Wangzhuang Road No. 1, Tsinghua Tongfang Technology Plaza, A block 18 layer Patentee after: Easyway Technology Co., Ltd. Address before: 100084 Haidian District 100084-82 mailbox Beijing Patentee before: Tsinghua University |