CN1410905A - Full distribution type aggregation network servicer system - Google Patents
Full distribution type aggregation network servicer system Download PDFInfo
- Publication number
- CN1410905A CN1410905A CN 02139288 CN02139288A CN1410905A CN 1410905 A CN1410905 A CN 1410905A CN 02139288 CN02139288 CN 02139288 CN 02139288 A CN02139288 A CN 02139288A CN 1410905 A CN1410905 A CN 1410905A
- Authority
- CN
- China
- Prior art keywords
- module
- bag
- delay
- packet
- cluster
- 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
Images
Abstract
The share medium type hub in the front end is the single entry point of the cluster. The requests of the clients through the hub are broadcast to the nodes in back end of the cluster. Each node in the cluster includes the isomorphic network card, the packet-filtering module, the delay balance module and the TCP stack-processing module. The network card in each node of the cluster is configured with the same IP address and MAC address. The packet-filtering module filters the data packets selectively. The delay balance module intercepts all request packets and delays the request packets. After the delayed time, the delay balance module submits the request packet to the TCP stack-processing module for processing. The TCP stack-processing module sends the data packets back to the client end.
Description
Technical field
The invention belongs to computer application field, be specifically related to a kind of full distributed cluster network service device system.
Background technology
It is the calculating period at center that the computer nowadays technology has entered with the network, and a large amount of application is all carried out round network, and the Performance And Reliability of server has been proposed more and more higher requirement.For example, along with the develop rapidly of Internet and user's severe increase, relatively more popular Web website can be because accessed number of times sharp increase and the request of process user in time, causes the user to wait for for a long time even goes whistle, and greatly reduces service quality.For the application of CPU intensity, such as have the Web service of CGI (Common Gateway Interface, CGI (Common gateway interface)) and database manipulation, the server performance bottleneck problem is then more outstanding.In addition, be applied in online popularization along with ecommerce etc. is key, the service disruption of any exception all will cause immeasurable loss, so the reliability of server is more and more important.
For the performance issue of settlement server, many companies and research institution have proposed extendible group system.These group systems are transferred to external network service request by a FEP real server of one group of rear end.Scheduling can realize in many places such as link layer, network layer or application layers with load-balancing technique.For example, EDDIE, Reverse-Proxy and pWEB use based on the method for application layer scheduling and set up a telescopic WEB cluster server.They are forwarded to different Web servers with the HTTP request that arrives, obtain the result after, return to the user again.The TCP Router of IBM and Network Dispatcher are the schedulers that is implemented in network layer.Linux Virtual Server is the colony dispatching software based on (SuSE) Linux OS.It provides multiple request scheduling scheme by the expansion IP stack to different network services.
Except software realizes that some companies have also proposed hardware solution, such as the LocalDirector of Cisco company.It carries out the rewriting that cluster server is imported and exported bag by hardware, and has obtained higher performance.It is said that it can dispatch hundreds thousand of TCP connection requests simultaneously, but this special hardware product price is very expensive, so that the general user is difficult to bear.
Above scheme has all adopted the centralized scheduling mode, though it has solved the overall performance bottleneck problem of cluster, all has common shortcoming, promptly seldom considers the reliability of FEP in this centralized scheduling.As the single entrance of whole cluster server, in a single day FEP breaks down, and whole server will paralyse, thereby may cause enormous economic loss.A simple solution is to be the ad hoc backup machine of this FEP, and when FEP broke down, backup machine replaced its IP address and works on.An important weak point of this fault-tolerant processing mode is that the TCP connection of having set up can all lose, and this can produce inconvenience even bring loss the user.In addition, above-mentioned various schemes based on centralized scheduling all are to connect as the load unit of measurement, so can not reach desirable load balancing effect.Especially provide in the isomeric group of multiple different services at each node, therefore the Internet resources that various services consume are different will become extremely unreasonable with the load that linking number is weighed each node.
In sum, there is following shortcoming in tradition based on the cluster server of centralized scheduling: 1, FEP becomes single failpoint; 2, will cause the load orientation out of true to be connected to load balance unit; 3, be directed on the same node owing to same IP, so dynamic port is dealt with improperly; 4, in using increasingly extensive end-to-end encryption application, FEP will become bottleneck point, thereby cause cluster relatively poor at encrypting the extensibility of using.
Summary of the invention
The present invention is directed to the shortcoming of centralized scheduling, a kind of full distributed cluster network service device system has been proposed, the notion of " FEP " in the centralized scheduling scheme has been eliminated by this system, has no single failpoint, no ink-bottle effect, is applicable to based on the service of dynamic protocol and the technique effect of fine-grained load balancing.
For achieving the above object, a kind of full distributed cluster network service device system is characterized in that:
The shared medium type hub of front end is as the single entrance of cluster, and the client is broadcast to its request on the node in the cluster of rear end by this hub;
Each node includes isomorphism network interface card, packet filtering module, delay balance module and TCP stack processing module in the cluster;
The isomorphism network interface card is used for the network interface card of inner each node of cluster is configured to identical IP address and MAC Address;
The packet filtering module is used for selectively filtering out the packet that the isomorphism network interface card transmits;
The delay balance module is responsible for intercepting all request package, and request package is postponed, and after finishing time delay, the delay balance module submits to TCP stack processing module to handle request package;
All packets that TCP stack processing module is beamed back send to gateway machine by true network interface card and deliver to client.
Described packet filtering module comprises packet handing module, PASS table, clocked flip-flop and rules administration module;
Described packet handing module resides at link layer, be responsible for when receiving the packet that the isomorphism network interface card transmits, mate with the rule of inquiry PASS table and to protocol number, source IP address, target ip address, source port, the target port of these bags, and finish corresponding processing action, this corresponding actions is divided three classes: abandon this bag, submit the bag intercepting submodule in the delay balance module to and submit kernel ICP/IP protocol stack processing module to;
Described PASS table is a rule set of describing with tuple, is responsible for the current packet handing module of definition and allows the packet that passes through;
Described clocked flip-flop is responsible at set intervals, and the triggering rule administration module carries out attended operation to the PASS table;
Described rules administration module is responsible for upgrading, delete the regularization term in the PASS table.
Described delay balance module comprises clocked flip-flop, load statistics module, time-delay computing module, bag interception module and bag queue table;
Described clocked flip-flop is responsible for triggering the load statistics module system resource is collected, and triggers the bag interception module simultaneously and obtain the packet of hanging up in the bag queue table;
Described load statistics module is responsible for collecting the situation of utilizing of various system resources, it and each devices communicating of system and the resource utilization of collecting each equipment, and these utilization factors are passed to the time-delay computing module as vector;
Described time-delay computing module calculates delay value according to the load vector that the load statistics module transmits;
Described bag interception module receives the SYN bag that transmits from packet handing module, and according to the delay value that the time-delay computing module transmits this SYN bag is suspended in the bag formation; In addition, when receiving the trigger pip that clocked flip-flop is sent, it will travel through the bag queue table, judge whether finish the time delay of each bag, if finish then fetch this datagram and send TCP stack processing module to;
Described bag queue table is used to deposit the SYN bag that is delayed, and each bag all carries a time-delay variable, is just fetched again by the bag interception module after finishing time delay.
The present invention has following technical characterstic:
The homogenous networks card technique is realized the single MAC reflection of the overall situation, be that the network interface card that receives services request in each node can both be bound a unified virtual ip address and virtual mac address, therefore can guarantee the services request bag that each node is all transmitted energy receiving front-end hub;
Based on the distributed balancing technique of time-delay competition, this technology can be carried out the time-delay of certain hour to all connection requests according to the load (cpu load, EMS memory occupation situation, magnetic disc i/o, network traffics etc.) of this locality, and load is heavy more, and it is long more to delay time;
The link layer packet filtering technology that state is arranged, can remember the data stream that belongs to same connection, allow those packets that connected to pass through, and can stop and connect and unfounded datagram is passed through, flow to the upper-layer protocol stack to avoid some nonsensical datagrams.
Based on above technical characterstic, the present invention has following technique effect: 1) do not have single failpoint
Because the aggregated server system that the present invention proposes adopts full distributed framework, so when wherein a node breaks down, all the other several nodes can continue adapter work, and can automatically be re-assigned to services request on the lightest node of load.2) no ink-bottle effect
In the centralized scheduling mode, when cluster provides when encrypt using, FEP needs the frequent bag to turnover to carry out being dispatched on the backend nodes after the encryption and decryption again, therefore can cause FEP to become bottleneck point and makes the extensibility variation of system.And distributed network scheduling scheme employing hub of the present invention replaces the front end dispatcher in the centralized scheduling, thereby has eliminated this ink-bottle effect.3) be applicable to service based on dynamic protocol
Dynamic protocol (as FTP, RTSP) is all adopted in present many services, because service node can be opened some new FPDP dynamically in the service process temporarily, in based on centralized scheduling, FEP need be known these transient ports and open them, so needing service node frequently communicates by letter with FEP, and only need upgrade the PASS table of this machine packet filtering module during the interim open port of service node in the distributed network of the present invention scheduling, thereby reduced network overhead.4) fine-grained load balancing
Weigh unit to be connected to load in the existing centralized scheduling scheme, therefore can cause the load of each service node unbalanced, and the scheduling mode that is based on load that the present invention adopts, can accurately locate the lightest node of load, reach splendid load balancing effect, so can be applied to the isomeric group system preferably.
Description of drawings
Fig. 1 is a kind of full distributed cluster network service device system architecture synoptic diagram of the present invention;
Fig. 2 is the workflow synoptic diagram of each node;
Fig. 3 is a packet filtering modular structure synoptic diagram among Fig. 1;
Delay balance modular structure synoptic diagram among Fig. 4 Fig. 1.
Embodiment
System architecture of the present invention as shown in Figure 1.The shared medium type hub of front end by means of the shared dielectric attribute of hub, makes the packet of all-access cluster to be received and to pass toward their link layers separately by the homogenous networks clamping of internal node as the single entrance of cluster.Inner all service nodes of cluster all are equipped with an isomorphism network interface card, and these network interface cards are all bound a public external IP address, to realize the single IP reflection of cluster.When client is sent out connection request, will produce a request package (SYN bag) as shown in Figure 1, the purpose IP address of SYN bag is the external IP address of cluster.When request package arrives the hub of cluster server front end, hub will be broadcast to this request package on the service node of rear end.
The workflow of each node as shown in Figure 2, the isomorphism network interface card of all nodes all will receive from the request package of hub broadcasting, and submit to the packet filtering module that resides in link layer, this module will allow all request package and those packets that connected to pass through, and refusal falls those packets that also do not connect.The packet filtering module is submitted to the upper strata IP stack to request package, the delay balance module that resides in the TCP layer will intercept all request package, and according to the load information of the load statistics submodule collection in this module request package is carried out the regular hour and postpone, being directly proportional with the loading condition of this machine time delay, (load is heavy more, it is long more to delay time), after finishing time delay, the delay balance module submits to TCP stack processing module to handle request package, sends back to request of client according to the treatment scheme of tcp protocol stack and confirms bag (SynAck bag).Therefore the request of sending of the lightest node of load is confirmed bag the earliest, and the request of all nodes at last confirms that bag will arrive client successively, i.e. the request of the node that load is light confirms that bag arrives at first, and the request of the heavy node of load is confirmed that bag is last and arrived.All packets that TCP stack processing module is beamed back all send to gateway machine by true network interface card.
As Fig. 1, the output data bag of all nodes all passes through this single exit point loopback of gateway machine and gives client.Because the characteristic of serving earlier first that tcp protocol stack is intrinsic, client will receive only the request affirmation bag (this request confirms that bag is sent by the lightest node of load) that arrives at first and feed back one and connect affirmation bag (ACK bag), and discard the bag of follow-up arrival and feed back a reset packet (RST bag).The bag that these clients are beamed back (ACK bag, a plurality of RST bags) by after all nodes receptions of cluster inside, each node will carry out certain processing: the node that receives the ACK bag will be finished establishment of connection, and in filtering module, adding a rule, the follow-up data bag that allows to belong to this connection passes through; The socket buffer that the node of reception RST bag is opened up release, promptly this node can not connect with client.At last, each client can be set up with the lightest node of load and is connected when cluster server conducts interviews, to finish fine-grained dynamic Network Load Balance.
Respectively each module is specified below.1, isomorphism network interface card
Isomorphism network interface card 4 adopts the homogenous networks card techniques to be realized, it is that network interface card with inner each node of cluster is configured to identical IP address (virtual ip address) and MAC Address (virtual mac address).This virtual ip address is exactly the open reference address of cluster server.
Each node all has an ARP high-speed cache.It has deposited nearest IP address to the map record between the hardware mac address.After a network packet process Route Selection, just can obtain the IP address of next stop router address or destination host, in the ARP high-speed cache, go to seek corresponding with it MAC Address then, if do not find, then ARP inquiry of broadcasting is wrapped in LAN (Local Area Network), this bag will be inquired about the MAC Address of the address correspondence of next stop router address or destination host, and the IP address that this is new and MAC Address mapping relations add in the arp cache of oneself, send the IP packet according to this MAC Address then.In present technique, for the consistance that guarantees on every node virtual ip address and virtual mac address mapping relations in the arp cache with avoid conflicting of IP address, need mask the response of ARP broadcast polling bag at kernel, implementation method is the code that inquiry is wrapped about ARP in the modification/net/ipv4/arp.c file.In addition, because the network interface card of each node all disposes same MAC Address and same IP address, the unreachable message of ICMP agreement udp port that can cause a certain node to send can cause network storm, in order to suppress this ICMP message storm, also to revise in the realization of present technique/ udp_recv () function among the net/ipv4/udp.c.
At last, virtual mac address with ifconfig command configuration network interface card, the MAC Address that disposes the network interface card of all nodes such as desire is virtual address 00:03:61:83:2a:8b, need carry out following several configuration step (the network interface card eth1 with node 1 is an example): the first step: second step of [root@node1/] #ifconfig eth1 down: the 3rd step of [root@node1/] #ifconfig eth1 ether HW 00:03:61:83:2a:8b: [root@node1/] #ifconfig eth1 up2 packet filtering module
Packet filtering module 5 adopts has the link layer packet filtering technology of state to be realized that it is used for selectively filtering out the packet that the isomorphism network interface card transmits.Because flowing into the packet of cluster all will be broadcast on each node through hub at every turn, so the isomorphism network interface card of each node will receive all these bags and submit to the kernel protocol stack on upper strata, to cause the kernel protocol stack of most of node can consume a large amount of resources like this and handle some irrelevant packets, particularly under the intensive situation of services request, the negative effect of this broadcasting will be more obvious.Therefore, the present invention adopts the packet filtering technology that state is arranged to abandon those irrelevant packets at link layer.The structural representation of this module as shown in Figure 4, specifically comprise: 1) packet handing module (10): this module resides at link layer, when receiving the packet that the isomorphism network interface card transmits, the rule that it will inquire about PASS table is mated protocol number, source IP address, target ip address, source port, the target port of these bags, and finish corresponding processing action, this corresponding actions is divided three classes: abandon this bag, submit the bag intercepting submodule (17) in the delay balance module to and submit kernel ICP/IP protocol stack processing module (7) to.2) PASS table (11): PASS table is a rule set of describing with tuple, and it has defined current packet handing module and has allowed the packet that passes through.In this table, all allow to pass through all about the packet of control linkage, as: SYN bag, SynAck bag, ACK bag, RST bag etc.; All packets that connected all will allow to pass through; Unless other packet system manager specifies, otherwise all acquiescence is discarded.3) clocked flip-flop (12): being responsible at set intervals, the triggering rule administration module carries out attended operation to the PASS table.4) rules administration module (13): be responsible for the regularization term in renewal, the deletion PASS table.When a TCP connects foundation or cancels, TCP stack processing module all will be notified this module, this module will be carrying out suitable processing then: when a connection was set up, it was converted into rule to link information (as: source IP, source port, Target IP, target port) and adds in the PASS table; When a connection is cancelled, it will delete corresponding rule in the PASS table.3 delay balance modules
Delay balance module 6 adopts the balancing technique of time-delay competition to be realized, it utilizes the three-way handshake mechanism of Transmission Control Protocol to finish the foundation that is connected of the lightest node of client and load.Work as client, because the natural response of protocol stack time-delay can not reflect the performance situation that system is all, institute thinks and produces a suitable delay value, needs the loading condition of collection all devices.The structural representation of this module specifically comprises as shown in Figure 3: 1) clocked flip-flop (14): be responsible for triggering the load statistics module system resource is collected, trigger the bag interception module simultaneously and obtain the packet of hanging up in the bag queue table.2) load statistics module (15): be responsible for to collect the situation of utilizing of various system resources, the resource utilization that it is communicated by letter and collect each equipment with each system equipment such as CPU, disk, internal memory and network interface card, and these utilization factors as vector (LOAD
CPU, LOAD
Disk, LOAD
Mem, LOAD
NIC) pass to the time-delay computing module.3) time-delay computing module (16): calculate suitable delay value according to the load vector that the load statistics module transmits, concrete computing formula is by providing as minor function.
Delay＝f(Load)＝RTT
MAX×(Load/Load
MAX)；
Load＝LOAD
CPU×Weight
CPU+Load
Disk×Weight
Disk+LOAD
Mem×Weight
Mem+
Load
NIC×Weight
NIC；
Wherein, RTT
MAXThe Round-Trip Time value that (middle Hop number is maximum) can reach when expression cluster and client communication farthest; Load
MAXBe load peak, this value is relevant with the concrete service that provides; Weight
CPU, Weight
Disk, Weight
MemAnd Weight
NICAll take normalized value, and different application or content employing are had different weights, and can combine the different definition of dynamic use with content scheduling, such as:
To static Web page, Weight
CPU=Weight
Mem=Weight
NIC=0, and Weight
Disk=1;
To dynamic web page, Weight
CPU=0.5, and Weight
Disk=0.5, Weight
Mem=Weight
NIC=0;
To Video service, Weight
CPU=Weight
Mem=Weight
NIC=Weight
Disk=0.25.4) bag interception module (17): the bag interception module receives the SYN bag that transmits from packet handing module (8), and according to the delay value that the time-delay computing module transmits this SYN bag is suspended in the bag formation; In addition, when receiving the trigger pip that clocked flip-flop is sent, it will travel through the bag queue table, judge whether finish the time delay of each bag, if finish then fetch this datagram and send TCP stack processing module (7) to.5) bag queue table (18): this table is deposited the SYN bag that is delayed, and each bag all carries a time-delay variable, is just fetched again by the bag interception module after finishing time delay.4TCP stack processing module
TCP stack processing module 7 adopts the tcp protocol stack treatment technology of safe enhancement mode to be realized, it is through the protocol module of revising, and has added some simple security mechanisms.Because potential broadcast overhead causes those " to fail to be elected " node easily and opens up too much socket memory headroom for the response three-way handshake, so group system suffers DoS attack easily.In this module, only need in the proc file system, open the syn-cookie technology.
The configuration of the hardware of each node and operating system can select for use configuration as shown in table 1 to come in addition specific implementation among the present invention, but is not limited to this configuration.
During concrete enforcement, external public network is linked the shared medium hub of cluster server outer end, links inner switch with a netting twine simultaneously.Then each node is configured to start the isomorphism network interface card accordingly.
| CPU | Internal memory | Mainboard | Operating system | Network |
| PIII866 | ??256M | Embedded two 100Mbps network interface unit | Linux kernel 2.4.5 | The 100M switch |
The software and hardware configuration of table 1 distributed self-adaption scheduling group system
During concrete enforcement, external public network is linked the shared medium hub of cluster server outer end, links inner switch with a netting twine simultaneously.Then each node is configured to start the isomorphism network interface card accordingly.
In conjunction with the accompanying drawings, as follows to the configuration instruction of total system: 1) PASS table
Which bag the PASS table is used for specifying to be received by this node, initial configuration is as shown in table 2, the implication of this table is: receive the packet that all zone bits are SYN, SynAck, RST, receive the bag that all send out from this node, and acquiescence discards other all bags.
Each field is explained as follows:
Protocol number: be divided into TCP, UDP, ICMP, ANY, ANY refers to any agreement;
Zone bit: be divided into SYN, SynAck, RST, ANY;
Source IP/ source port: the source IP address/source port of packet;
Purpose IP/ destination interface: the purpose IP address/destination interface of packet;
Measure: refer to the coupling packet with which kind of measure of access, be divided into two kinds of PASS (reception) and DROP (refusal).
| Protocol number | Zone bit | Source IP | Source port | Purpose IP | Destination interface | Measure |
| TCP | ?SYN | ??ANY | ??ANY | ??17.0.0.2 | ????21 | ??PASS |
| TCP | ?SynAck | ??ANY | ??ANY | ??17.0.0.2 | ????21 | ??PASS |
| TCP | ?RST | ??ANY | ??ANY | ??17.0.0.2 | ????21 | ??PASS |
| ANY | ?ANY | ??17.0.0.2 | ??ANY | ???ANY | ????ANY | ??PASS |
| ANY | ?ANY | ??ANY | ??ANY | ???ANY | ????ANY | ??DROP |
Table 2 provides the FTP node configuration example (establishing this node IP is 17.0.0.2) 2 of service) the time-delay computing module
The time-delay computing module need calculate corresponding time-delay size according to the load weights of COS and CPU, internal memory, network interface card and disk, and corresponding relation is as shown in table 3.
| COS | ??Weight CPU | ??Weight Mem | ??Weight NIC | ??Weight Disk |
| ??HTTP | ??0.5 | ???0.5 | ???0 | ????1 |
| ??FTP | ??0 | ???0.5 | ???0 | ????0.5 |
| ??RSTP | ??0.25 | ???0.25 | ???0.25 | ????0.25 |
The mapping table of table 3 COS and load weights
Claims (3)
1, a kind of full distributed cluster network service device system is characterized in that:
The shared medium type hub of front end is as the single entrance of cluster, and the client is broadcast to its request on the node in the cluster of rear end by this hub;
Each node includes isomorphism network interface card, packet filtering module, delay balance module and TCP stack processing module in the cluster;
The isomorphism network interface card is used for the network interface card of inner each node of cluster is configured to identical IP address and MAC Address;
The packet filtering module is used for selectively filtering out the packet that the isomorphism network interface card transmits;
The delay balance module is responsible for intercepting all request package, and request package is postponed, and after finishing time delay, the delay balance module submits to TCP stack processing module to handle request package;
All packets that TCP stack processing module is beamed back send to gateway machine by true network interface card and deliver to client.
2, system according to claim 1 is characterized in that:
Described packet filtering module comprises packet handing module, PASS table, clocked flip-flop and rules administration module;
Described packet handing module resides at link layer, be responsible for when receiving the packet that the isomorphism network interface card transmits, mate with the rule of inquiry PASS table and to protocol number, source IP address, target ip address, source port, the target port of these bags, and finish corresponding processing action, this corresponding actions is divided three classes: abandon this bag, submit the bag intercepting submodule in the delay balance module to and submit kernel ICP/IP protocol stack processing module to;
Described PASS table is a rule set of describing with tuple, is responsible for the current packet handing module of definition and allows the packet that passes through;
Described clocked flip-flop is responsible at set intervals, and the triggering rule administration module carries out attended operation to the PASS table;
Described rules administration module is responsible for upgrading, delete the regularization term in the PASS table.
3, system according to claim 1 and 2 is characterized in that:
Described delay balance module comprises clocked flip-flop, load statistics module, time-delay computing module, bag interception module and bag queue table;
Described clocked flip-flop is responsible for triggering the load statistics module system resource is collected, and triggers the bag interception module simultaneously and obtain the packet of hanging up in the bag queue table;
Described load statistics module is responsible for collecting the situation of utilizing of various system resources, it and each devices communicating of system and the resource utilization of collecting each equipment, and these utilization factors are passed to the time-delay computing module as vector;
Described time-delay computing module calculates delay value according to the load vector that the load statistics module transmits;
Described bag interception module receives the SYN bag that transmits from packet handing module, and according to the delay value that the time-delay computing module transmits this SYN bag is suspended in the bag formation; In addition, when receiving the trigger pip that clocked flip-flop is sent, it will travel through the bag queue table, judge whether finish the time delay of each bag, if finish then fetch this datagram and send TCP stack processing module to;
Described bag queue table is used to deposit the SYN bag that is delayed, and each bag all carries a time-delay variable, is just fetched again by the bag interception module after finishing time delay.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 02139288 CN1206600C (en) | 2002-11-14 | 2002-11-14 | Full distribution type aggregation network servicer system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 02139288 CN1206600C (en) | 2002-11-14 | 2002-11-14 | Full distribution type aggregation network servicer system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1410905A true CN1410905A (en) | 2003-04-16 |
| CN1206600C CN1206600C (en) | 2005-06-15 |
Family
ID=4750007
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 02139288 Expired - Fee Related CN1206600C (en) | 2002-11-14 | 2002-11-14 | Full distribution type aggregation network servicer system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN1206600C (en) |
Cited By (29)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1315298C (en) * | 2003-07-01 | 2007-05-09 | 智邦科技股份有限公司 | System and method for synchronous packaging processing |
| CN100430917C (en) * | 2005-04-22 | 2008-11-05 | 天津曙光计算机产业有限公司 | Method for self-constructing group operating system core and intelligent constructor |
| CN100502271C (en) * | 2003-09-17 | 2009-06-17 | 华为技术有限公司 | Calling admission control method of WCDMA access system |
| CN101924698A (en) * | 2010-07-22 | 2010-12-22 | 福建星网锐捷网络有限公司 | Method, system and equipment for balancing two-layer domain load based on IP unicast route |
| US7978714B2 (en) | 2004-07-23 | 2011-07-12 | Citrix Systems, Inc. | Methods and systems for securing access to private networks using encryption and authentication technology built in to peripheral devices |
| US7978716B2 (en) | 2003-11-24 | 2011-07-12 | Citrix Systems, Inc. | Systems and methods for providing a VPN solution |
| CN102143031A (en) * | 2011-04-02 | 2011-08-03 | 中兴通讯股份有限公司 | Network access seamless switching method and terminal |
| US8014421B2 (en) | 2004-07-23 | 2011-09-06 | Citrix Systems, Inc. | Systems and methods for adjusting the maximum transmission unit by an intermediary device |
| CN102238140A (en) * | 2010-04-27 | 2011-11-09 | 鸿富锦精密工业（深圳）有限公司 | System and method for sharing server by clients |
| CN102316539A (en) * | 2006-09-28 | 2012-01-11 | 高通股份有限公司 | Be used for the predictive qos resource allocation that fast session is set up |
| US8255456B2 (en) | 2005-12-30 | 2012-08-28 | Citrix Systems, Inc. | System and method for performing flash caching of dynamically generated objects in a data communication network |
| US8261057B2 (en) | 2004-06-30 | 2012-09-04 | Citrix Systems, Inc. | System and method for establishing a virtual private network |
| US8301839B2 (en) | 2005-12-30 | 2012-10-30 | Citrix Systems, Inc. | System and method for performing granular invalidation of cached dynamically generated objects in a data communication network |
| WO2013020459A1 (en) * | 2011-08-10 | 2013-02-14 | 华为技术有限公司 | Distributed cluster processing system and message processing method thereof |
| CN103139247A (en) * | 2011-11-28 | 2013-06-05 | 百度在线网络技术（北京）有限公司 | Keep-alive method, system and device for rear-end server in load balancing system |
| CN101394400B (en) * | 2007-05-18 | 2013-07-10 | 辉达公司 | Method for establishing network connection for avoiding damage of ARP cache |
| US8495305B2 (en) | 2004-06-30 | 2013-07-23 | Citrix Systems, Inc. | Method and device for performing caching of dynamically generated objects in a data communication network |
| US8499057B2 (en) | 2005-12-30 | 2013-07-30 | Citrix Systems, Inc | System and method for performing flash crowd caching of dynamically generated objects in a data communication network |
| US8549149B2 (en) | 2004-12-30 | 2013-10-01 | Citrix Systems, Inc. | Systems and methods for providing client-side accelerated access to remote applications via TCP multiplexing |
| US8700695B2 (en) | 2004-12-30 | 2014-04-15 | Citrix Systems, Inc. | Systems and methods for providing client-side accelerated access to remote applications via TCP pooling |
| US8706877B2 (en) | 2004-12-30 | 2014-04-22 | Citrix Systems, Inc. | Systems and methods for providing client-side dynamic redirection to bypass an intermediary |
| US8739274B2 (en) | 2004-06-30 | 2014-05-27 | Citrix Systems, Inc. | Method and device for performing integrated caching in a data communication network |
| CN103929365A (en) * | 2014-03-25 | 2014-07-16 | 上海格尔软件股份有限公司 | Load balancing system and method suitable for a plurality of UDP services |
| US8856777B2 (en) | 2004-12-30 | 2014-10-07 | Citrix Systems, Inc. | Systems and methods for automatic installation and execution of a client-side acceleration program |
| CN104283943A (en) * | 2014-09-22 | 2015-01-14 | 珠海许继芝电网自动化有限公司 | Communication optimizing method for cluster server |
| US8954595B2 (en) | 2004-12-30 | 2015-02-10 | Citrix Systems, Inc. | Systems and methods for providing client-side accelerated access to remote applications via TCP buffering |
| CN105812425A (en) * | 2014-12-30 | 2016-07-27 | Tcl集团股份有限公司 | Method for establishing cross-room server cluster and node servers |
| CN107087014A (en) * | 2017-01-24 | 2017-08-22 | 无锡英威腾电梯控制技术有限公司 | A kind of load-balancing method and its controller |
| CN109376123A (en) * | 2014-08-12 | 2019-02-22 | 华为技术有限公司 | Manage method, distributed memory system and the management node of file |
-
2002
- 2002-11-14 CN CN 02139288 patent/CN1206600C/en not_active Expired - Fee Related
Cited By (53)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1315298C (en) * | 2003-07-01 | 2007-05-09 | 智邦科技股份有限公司 | System and method for synchronous packaging processing |
| CN100502271C (en) * | 2003-09-17 | 2009-06-17 | 华为技术有限公司 | Calling admission control method of WCDMA access system |
| US8559449B2 (en) | 2003-11-11 | 2013-10-15 | Citrix Systems, Inc. | Systems and methods for providing a VPN solution |
| US7978716B2 (en) | 2003-11-24 | 2011-07-12 | Citrix Systems, Inc. | Systems and methods for providing a VPN solution |
| US8495305B2 (en) | 2004-06-30 | 2013-07-23 | Citrix Systems, Inc. | Method and device for performing caching of dynamically generated objects in a data communication network |
| US8726006B2 (en) | 2004-06-30 | 2014-05-13 | Citrix Systems, Inc. | System and method for establishing a virtual private network |
| US8261057B2 (en) | 2004-06-30 | 2012-09-04 | Citrix Systems, Inc. | System and method for establishing a virtual private network |
| US8739274B2 (en) | 2004-06-30 | 2014-05-27 | Citrix Systems, Inc. | Method and device for performing integrated caching in a data communication network |
| US8897299B2 (en) | 2004-07-23 | 2014-11-25 | Citrix Systems, Inc. | Method and systems for routing packets from a gateway to an endpoint |
| US8046830B2 (en) | 2004-07-23 | 2011-10-25 | Citrix Systems, Inc. | Systems and methods for network disruption shielding techniques |
| US8019868B2 (en) | 2004-07-23 | 2011-09-13 | Citrix Systems, Inc. | Method and systems for routing packets from an endpoint to a gateway |
| US8892778B2 (en) | 2004-07-23 | 2014-11-18 | Citrix Systems, Inc. | Method and systems for securing remote access to private networks |
| US8014421B2 (en) | 2004-07-23 | 2011-09-06 | Citrix Systems, Inc. | Systems and methods for adjusting the maximum transmission unit by an intermediary device |
| US8914522B2 (en) | 2004-07-23 | 2014-12-16 | Citrix Systems, Inc. | Systems and methods for facilitating a peer to peer route via a gateway |
| US7978714B2 (en) | 2004-07-23 | 2011-07-12 | Citrix Systems, Inc. | Methods and systems for securing access to private networks using encryption and authentication technology built in to peripheral devices |
| US8291119B2 (en) | 2004-07-23 | 2012-10-16 | Citrix Systems, Inc. | Method and systems for securing remote access to private networks |
| US8634420B2 (en) | 2004-07-23 | 2014-01-21 | Citrix Systems, Inc. | Systems and methods for communicating a lossy protocol via a lossless protocol |
| US9219579B2 (en) | 2004-07-23 | 2015-12-22 | Citrix Systems, Inc. | Systems and methods for client-side application-aware prioritization of network communications |
| US8351333B2 (en) | 2004-07-23 | 2013-01-08 | Citrix Systems, Inc. | Systems and methods for communicating a lossy protocol via a lossless protocol using false acknowledgements |
| US8363650B2 (en) | 2004-07-23 | 2013-01-29 | Citrix Systems, Inc. | Method and systems for routing packets from a gateway to an endpoint |
| US8954595B2 (en) | 2004-12-30 | 2015-02-10 | Citrix Systems, Inc. | Systems and methods for providing client-side accelerated access to remote applications via TCP buffering |
| US8700695B2 (en) | 2004-12-30 | 2014-04-15 | Citrix Systems, Inc. | Systems and methods for providing client-side accelerated access to remote applications via TCP pooling |
| US8856777B2 (en) | 2004-12-30 | 2014-10-07 | Citrix Systems, Inc. | Systems and methods for automatic installation and execution of a client-side acceleration program |
| US8706877B2 (en) | 2004-12-30 | 2014-04-22 | Citrix Systems, Inc. | Systems and methods for providing client-side dynamic redirection to bypass an intermediary |
| US8549149B2 (en) | 2004-12-30 | 2013-10-01 | Citrix Systems, Inc. | Systems and methods for providing client-side accelerated access to remote applications via TCP multiplexing |
| US8788581B2 (en) | 2005-01-24 | 2014-07-22 | Citrix Systems, Inc. | Method and device for performing caching of dynamically generated objects in a data communication network |
| US8848710B2 (en) | 2005-01-24 | 2014-09-30 | Citrix Systems, Inc. | System and method for performing flash caching of dynamically generated objects in a data communication network |
| CN100430917C (en) * | 2005-04-22 | 2008-11-05 | 天津曙光计算机产业有限公司 | Method for self-constructing group operating system core and intelligent constructor |
| US8301839B2 (en) | 2005-12-30 | 2012-10-30 | Citrix Systems, Inc. | System and method for performing granular invalidation of cached dynamically generated objects in a data communication network |
| US8499057B2 (en) | 2005-12-30 | 2013-07-30 | Citrix Systems, Inc | System and method for performing flash crowd caching of dynamically generated objects in a data communication network |
| US8255456B2 (en) | 2005-12-30 | 2012-08-28 | Citrix Systems, Inc. | System and method for performing flash caching of dynamically generated objects in a data communication network |
| CN102316539A (en) * | 2006-09-28 | 2012-01-11 | 高通股份有限公司 | Be used for the predictive qos resource allocation that fast session is set up |
| US9253092B2 (en) | 2006-09-28 | 2016-02-02 | Qualcomm Incorporated | Predictive QoS resource allocation for rapid session establishment |
| CN101394400B (en) * | 2007-05-18 | 2013-07-10 | 辉达公司 | Method for establishing network connection for avoiding damage of ARP cache |
| CN102238140A (en) * | 2010-04-27 | 2011-11-09 | 鸿富锦精密工业（深圳）有限公司 | System and method for sharing server by clients |
| CN101924698A (en) * | 2010-07-22 | 2010-12-22 | 福建星网锐捷网络有限公司 | Method, system and equipment for balancing two-layer domain load based on IP unicast route |
| CN101924698B (en) * | 2010-07-22 | 2012-12-12 | 福建星网锐捷网络有限公司 | Method, system and equipment for balancing two-layer domain load based on IP unicast route |
| WO2012136053A1 (en) * | 2011-04-02 | 2012-10-11 | 中兴通讯股份有限公司 | Method and system for network access seamless switching |
| CN102143031A (en) * | 2011-04-02 | 2011-08-03 | 中兴通讯股份有限公司 | Network access seamless switching method and terminal |
| CN102143031B (en) * | 2011-04-02 | 2016-03-02 | 中兴通讯股份有限公司 | A kind of network access seamless switching method and terminal |
| US10404773B2 (en) | 2011-08-10 | 2019-09-03 | Huawei Technologies Co., Ltd. | Distributed cluster processing system and packet processing method thereof |
| WO2013020459A1 (en) * | 2011-08-10 | 2013-02-14 | 华为技术有限公司 | Distributed cluster processing system and message processing method thereof |
| CN103139247A (en) * | 2011-11-28 | 2013-06-05 | 百度在线网络技术（北京）有限公司 | Keep-alive method, system and device for rear-end server in load balancing system |
| CN103139247B (en) * | 2011-11-28 | 2015-11-25 | 百度在线网络技术（北京）有限公司 | The keepalive method of back-end server, system and device in SiteServer LBS |
| CN103929365A (en) * | 2014-03-25 | 2014-07-16 | 上海格尔软件股份有限公司 | Load balancing system and method suitable for a plurality of UDP services |
| CN103929365B (en) * | 2014-03-25 | 2019-05-14 | 格尔软件股份有限公司 | A kind of SiteServer LBS and method suitable for UDP service |
| CN109376123A (en) * | 2014-08-12 | 2019-02-22 | 华为技术有限公司 | Manage method, distributed memory system and the management node of file |
| US11656763B2 (en) | 2014-08-12 | 2023-05-23 | Huawei Technologies Co., Ltd. | File management method, distributed storage system, and management node |
| CN104283943A (en) * | 2014-09-22 | 2015-01-14 | 珠海许继芝电网自动化有限公司 | Communication optimizing method for cluster server |
| CN105812425B (en) * | 2014-12-30 | 2019-04-19 | Tcl集团股份有限公司 | A kind of method and node server of the establishment across server in machine room cluster |
| CN105812425A (en) * | 2014-12-30 | 2016-07-27 | Tcl集团股份有限公司 | Method for establishing cross-room server cluster and node servers |
| CN107087014A (en) * | 2017-01-24 | 2017-08-22 | 无锡英威腾电梯控制技术有限公司 | A kind of load-balancing method and its controller |
| CN107087014B (en) * | 2017-01-24 | 2020-12-15 | 无锡英威腾电梯控制技术有限公司 | Load balancing method and controller thereof |
Also Published As
| Publication number | Publication date |
|---|---|
| CN1206600C (en) | 2005-06-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN1206600C (en) | Full distribution type aggregation network servicer system | |
| US6665304B2 (en) | Method and apparatus for providing an integrated cluster alias address | |
| CN101217493B (en) | TCP data package transmission method | |
| US7684423B2 (en) | System and method for virtual network interface cards based on internet protocol addresses | |
| US7742474B2 (en) | Virtual network interface cards with VLAN functionality | |
| US7792140B2 (en) | Reflecting the bandwidth assigned to a virtual network interface card through its link speed | |
| US7733890B1 (en) | Network interface card resource mapping to virtual network interface cards | |
| US20100054123A1 (en) | Method and device for hign utilization and efficient flow control over networks with long transmission latency | |
| EP1063818A2 (en) | System for multi-layer provisioning in computer networks | |
| US20040260745A1 (en) | Load balancer performance using affinity modification | |
| CN1875585A (en) | Dynamic unknown L2 flooding control with MAC limits | |
| CN109698796A (en) | A kind of high performance network SiteServer LBS and its implementation | |
| CN103049336A (en) | Hash-based network card soft interrupt and load balancing method | |
| US8630296B2 (en) | Shared and separate network stack instances | |
| CN1284094C (en) | Distributed parallel scheduling wide band network server system | |
| US7469295B1 (en) | Modified round robin load balancing technique based on IP identifier | |
| Chen et al. | A survey of kernel-bypass techniques in network stack | |
| CN102158406B (en) | Intelligent routing method for computer network links | |
| WO2021120633A1 (en) | Load balancing method and related device | |
| CN101867521A (en) | Multilink accessing and flow load dispatching managing method | |
| US9491098B1 (en) | Transparent network multipath utilization through encapsulation | |
| US7591011B1 (en) | Assigning higher priority to transactions based on subscription level | |
| Ke et al. | Load balancing using P4 in software-defined networks | |
| Liu et al. | A Modified Round-Robin load balancing algorithm based on content of request | |
| Goldszmidt | Load management for scaling up Internet services |
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 | ||
| C19 | Lapse of patent right due to non-payment of the annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |