Movatterモバイル変換


[0]ホーム

URL:


CN109412866A - A kind of active detecting method of multi-tenant cloud platform security isolation - Google Patents

A kind of active detecting method of multi-tenant cloud platform security isolation
Download PDF

Info

Publication number
CN109412866A
CN109412866ACN201811473272.8ACN201811473272ACN109412866ACN 109412866 ACN109412866 ACN 109412866ACN 201811473272 ACN201811473272 ACN 201811473272ACN 109412866 ACN109412866 ACN 109412866A
Authority
CN
China
Prior art keywords
node
tenant
detection
security isolation
cloud platform
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
Application number
CN201811473272.8A
Other languages
Chinese (zh)
Other versions
CN109412866B (en
Inventor
王利明
葛思江
李兆璨
孔同
杨倩
马多贺
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Institute of Information Engineering of CAS
Original Assignee
Institute of Information Engineering of CAS
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Institute of Information Engineering of CASfiledCriticalInstitute of Information Engineering of CAS
Priority to CN201811473272.8ApriorityCriticalpatent/CN109412866B/en
Publication of CN109412866ApublicationCriticalpatent/CN109412866A/en
Application grantedgrantedCritical
Publication of CN109412866BpublicationCriticalpatent/CN109412866B/en
Activelegal-statusCriticalCurrent
Anticipated expirationlegal-statusCritical

Links

Classifications

Landscapes

Abstract

Translated fromChinese

本发明公开了一种多租户云平台安全隔离的主动检测方法,实现在运行时对多租户云平台的安全隔离性进行实时检测。方法步骤包括:(1)定义图模型;(2)定义检测基准;(3)初始状态生成;(4)初始隔离检测;(5)运行时隔离检测。本发明方法以图的形式对多租户云平台的配置状态进行组织建模,并建立了基于租户安全隔离的检测基准和基于用户组安全隔离的检测基准,然后实施安全隔离检测,以及时发现潜在隔离失效威胁对云内租户造成的危害。本方法具有可扩展性强、时效性高的特点,解决了现有多租户云平台中的安全隔离检测问题,提高了多租户云平台的安全能力,从而满足现有应用场景的需求。

The invention discloses an active detection method for the security isolation of a multi-tenant cloud platform, which realizes real-time detection of the security isolation of the multi-tenant cloud platform during operation. The method steps include: (1) defining a graph model; (2) defining a detection benchmark; (3) initial state generation; (4) initial isolation detection; (5) runtime isolation detection. The method of the invention organizes and models the configuration state of the multi-tenant cloud platform in the form of a graph, establishes a detection benchmark based on tenant security isolation and a detection benchmark based on user group security isolation, and then implements security isolation detection to discover potential potential Isolate failure threats to in-cloud tenants. The method has the characteristics of strong scalability and high timeliness, solves the problem of security isolation detection in the existing multi-tenant cloud platform, improves the security capability of the multi-tenant cloud platform, and thus meets the needs of existing application scenarios.

Description

A kind of active detecting method of multi-tenant cloud platform security isolation
Technical field
The invention belongs to a kind of detection sides of security isolation in cloud security technical field more particularly to multi-tenant cloud platformMethod.
Background technique
Cloud has been increasingly becoming the development trend of IT infrastructure, quilt as one of technology with the fastest developing speed in recent yearsIt is more and more widely used.It is by proposing a kind of resource (including computing resource, Internet resources, storage resource etc.) Chi HuajiArt allows infrastructure by the shared use of multiple tenants, to realize a kind of convenient, on-demand method of service.
However, multi-tenant technology has broken the barrier between physical equipment in cloud platform, the utilization rate of resource is being improvedMeanwhile also inevitably bringing severe safety problem.On the one hand, cloud service provider need to provide corresponding tenant everyThe system of disembarking ensures specific resources not by unauthorized access;On the other hand, if isolated failure in shared environment, malice tenant in cloudBreak isolation mech isolation test, initiates unauthorized access, the data assets for making other tenants are faced into huge security risk.
Therefore, in multi-tenant cloud platform, security isolation is particularly important, and how to improve the security capabilities of cloud platform,It is a technical problem that needs to be urgently solved by technical personnel in the field at present.Currently, aiming at the problem that security isolation, correlation solutionCertainly scheme includes three categories: defence, subsequent retrospect and run-time check in advance.
The scheme defendd in advance, which mainly passes through, reinforces the means such as access control, enhancing security module, the energy of limitation access in advancePower.By the retrieval discovery to existing patent, Chinese patent literature CN104580505A discloses (bulletin) day 2015.04.29,Disclose a kind of tenant's partition method and virtual switch, comprising: virtual switch is each virtual machine point on each physical hostWith corresponding for identifying the virtual local area network tags of tenant's message;Virtual switch is that each tenant distributes corresponding virtual tenantNetwork (VTN) identifier, generates the corresponding virtual network of each tenant.
The scheme traced afterwards is chased after by the way that the data of the acquisition in a period of time are excavated and are associated in ex-post analysisIt traces back in cloud after already present attack path, enforcing remedies measure.By the retrieval discovery to existing patent, Chinese patent literature numberCN107566369A discloses (bulletin) day 2018.01.09, discloses a kind of for industry control infrastructure progress security isolation and anti-Imperial efficiency evaluation method, this method can equally act on IT infrastructure based on cloud.It include: to establish an isolation and preventImperial technology model;Information is collected based on technology model;Information carries out algorithm analysis according to the collected data;Pass through analysisThe ability of various attacks is resisted when operation system is under attack, then assesses the effectiveness of security isolation measure.
Although essential a part in the above two classes solution and cloud security protection system, its deficiency existsIn the former can not cope with the fortuitous event occurred in cloud platform operational process, such as loophole, error configurations;The latter can only be to cloudThe security threat inside occurred is analyzed and is remedied, and preventive effect can not be played.
Third class scheme may make even if cloud platform just during operation by being checked at runtime, can also be withThe potential security isolation of the discovery of active threatens, to make up the deficiency of first two scheme.The existing safety for tenant's rankIsolation real-time detection method includes: (1) by token network data packet, realizes real-time tracing and the analysis of convection current.But this methodIt cannot check that potential failure threatens;(2) by solving Boolean satisfiability, network in cloud is verified.This method is depositedIt is small to store up expense, but real-time is poor.(3) by carrying out information collection to cloud platform, graph model is established, real-time update and mould are implementedType analysis.This respect work is concentrated mainly on the configuration compliance according to provider API detection cloud infrastructure.This method is real-timeProperty it is strong, but for multi-tenant cloud platform security isolation problem it is further proposed that prolongable detection scheme.
The prior art has that expansibility is weaker, timeliness is not strong.The present invention proposes a kind of security isolation detectionMethod can detect potential isolated failure threat that may be present in the cloud environment of dynamic change, to make up existing in timeThe deficiency of method improves the security capabilities of multi-tenant cloud platform.
Summary of the invention
Present invention aim to address the deficiencies of existing multi-tenant cloud platform security isolation technology, it is therefore proposed that a kind of rightThe method that configuration carries out active safety detection in cloud finds isolated failure threat that may be present in multi-tenant cloud platform.The partyMethod has the characteristics that scalability is strong, timeliness is high.
To achieve the above object, the present invention takes following methods, comprising:
(1) graph model is defined: Directed Graph Model G=(V, E, C), V of the definition comprising each node in multi-tenant cloud platformIndicate the cluster of the node of be provided with service in cloud platform, any node meets v in clusteri∈V;E is indicated between each nodeSide, there are any bar side e in clusteri,jIndicate node viIt may have access to another node vj, ei,j∈E;C is indicated between nodeConnected relation, ci,jIndicate node viWith vjBetween connected relation, ci,j∈C.I, j indicate the mark of two nodes.
When constructing Directed Graph Model, if ci,j=1, then it is assumed that v in figure GiWith vjBetween there are side ei,jIf ci,j=0, then viWith vjBetween be not present side ei,j.Above-mentioned Directed Graph Model G is stored in the form of two-dimentional adjacency matrix, is denoted as:
(2) definition detection benchmark: in multi-tenant cloud platform, there are multiple tenants to belong to different rents by Admin AdministrationUnauthorized access can not be initiated between the node at family;Inside each tenant, there are multiple user groups, by administrator or tenant administratorManagement, the node for belonging to different user groups can not initiate unauthorized access.Therefore, security isolation includes two class scenes: first is that differentSecurity isolation between tenant need to verify the potential threat that whether there is unauthorized access between each tenant;Second is that same tenantSecurity isolation between interior different user groups need to be verified and whether there is the latent of unauthorized access in same tenant between different user groupsIt is threatening.For the above scene, detection benchmark based on tenant's security isolation is defined respectively and based on user group security isolationDetect benchmark.
(21) based on the detection benchmark of tenant's security isolation: when attacker obtains in administrator right or multi-tenant cloud platformThere are when malice administrator, the configuration of multi-tenant cloud platform can be maliciously tampered in portion, make to originally belong to different tenants node itBetween can initiate unauthorized access, tenant's security isolation is destroyed.Therefore, the correct mapping relationship f for establishing tenant and node, asThe detection benchmark of tenant's security isolation.If n indicates the mark of node, tenentnIndicate the mark for being identified as the node tenant of nKnow, mapping relations be denoted as:
F (n)=tenentn
(22) based on the detection benchmark of user group security isolation: for the user for belonging to same tenant in multi-tenant cloud platformGroup, definition L are user group grade, and L value can be uppr, normal or lower.L value may have access to non-for the user group of upprAll user groups of uppr, can not be by other users group access;L value be normal user group can be all with L value normal andMark value is all that the user group of x is exchanged visits, and x is the integer greater than zero;L value be lower user group can by all user's group access,It can not initiate to access.For the node in user group, if n indicates the mark of node, levelnIndicate the mark for being identified as the node of nNote value, then the mark value level of node nnIs defined as:
When attacker obtains inside administrator or tenant's administrator right and multi-tenant cloud platform, there are malice administratorsOr when malice tenant administrator, the configuration of the grade of user group can be maliciously tampered, make different grades of user group node itBetween can initiate unauthorized access, user group security isolation is destroyed.Therefore, each user group and node are established in same tenant justMapping relations are denoted as by true mapping relations y as the detection benchmark of user group security isolation:
Y (n)=leveln
(3) original state generates: acquiring the configuration data of all nodes in multi-tenant cloud platform under original state.According to stepSuddenly Directed Graph Model G=(V, E, C) defined in (1) traverses configuration data and records current traverse node viWith other sectionsPoint vjConnected relation ci,j∈C.If viWith vjIt is connected to, then ci,j=1, otherwise ci,j=0, then with step (1) the definition structureBuild the graph model G based on original stateinit
(4) benchmark, digraph G described in detecting step (3) initial isolation detection: are detected described in foundation step (2)initWhether security isolation is met, and detection process includes initialization vertex ticks and initialization two stages of nodal test:
(41) vertex ticks process is initialized are as follows: successively traversal step (3) the digraph GinitIn all nodes.TimeIt is node to be detected by current accessed vertex ticks, and be inserted into queue Detect to be detected from tail end during going through.
(42) nodal test process is initialized are as follows: be successively read node up to queue is from queue Detect head end to be detectedThe current node that reads is denoted as v by skydet.Detect vdetWhether security isolation demand is met, and upon completion of the assays by vdetFrom the beginningEnd removes queue Detect to be detected.Specific detection content includes: the tenant's security isolation inspection executed based on detection benchmark (21)It surveys, the user group security isolation detection executed based on detection benchmark (22).
(5) isolation detection when running: being denoted as time for the current time of running, if at the time moment, administrator or tenant's pipeReason person, which implements configuration operation to cloud platform, causes digraph state that will update, then current digraph is denoted as Gtime, and according toBenchmark is detected according to described in step (2), detects G when updatingtimeWhether security isolation demand is met, and detection process includes updatingWhen vertex ticks and update when two stages of nodal test:
(51) update when node labeling process are as follows: described in step (5) configuration operation include creation example, delete example,It creates security strategy, delete security strategy.When in cloud platform administrator or tenant administrator implement aforesaid operations when, increment is moreNew graph model Gtime, and will be updated vertex ticks is node to be detected, is inserted into queue Detect to be detected from tail end.
(52) node detection process when updating are as follows: be successively read node up to queue is from queue Detect head end to be detectedThe current node that reads is denoted as v by skydet.Detect vdetWhether security isolation demand is met, and upon completion of the assays by vdetFrom the beginningEnd removes queue Detect to be detected.Specific detection content includes: the tenant's security isolation inspection executed based on detection benchmark (21)It surveys, the user group security isolation detection executed based on detection benchmark (22).
Further, according to the active detecting method of the multi-tenant cloud platform security isolation, it is characterised in that: step(51) incremental update described in, when implementation, specifically may include following four classes situation:
Side insertion: make step (5) the graph model G when creating or deleting new security strategytimeIn node viIt may have access tovjWhen, connected relation c described in setting steps (1)i,j=1, make adjacency matrix A [i] [j]=1, thereby executing side insertion.
Edge contract: make step (5) the graph model G when creating or deleting new security strategytimeIn node viIt can not visitAsk vjWhen, connected relation c described in setting steps (1)i,j=0, make adjacency matrix A [i] [j]=0, thereby executing edge contract.
Point insertion: when creating new example, in step (5) the graph model GtimeMiddle insertion new node, is denoted as vnew, newIndicate node identification.If remembering GtimeIn other nodes be identified as k, adjacency matrix A [new] [k] described in initialization step (1)=0 and A [new] [k]=0.As node vnewIt may have access to vkWhen, connected relation c described in setting steps (1)new,k=1, make to abutMatrix A [new] [k]=1;As node vkIt may have access to vnewWhen, connected relation c is setk,new=1, make adjacency matrix A [k] [new]=1.Thereby executing an insertion.
Point deletion: when deleting new example, in step (5) the graph model GtimeThe corresponding node of middle positioning example,It is denoted as vdel, del expression node identification.At this time by the way that del=-1 is arranged by node vdelLabeled as failure, it is not involved in detection.FromAnd execute point deletion.
Further, according to the active detecting method of the multi-tenant cloud platform security isolation, it is characterised in that: step(42), tenant's security isolation described in (52) detects, and is according to the detection base based on tenant's security isolation described in step (2)Standard, the node v that queue Detect head end to be detected described in detecting step (42), (52) is readdetWhether tenant safety is metIsolation.
For node vdetSpecific detection process include: to scheme G described in traversal step (3)initOr described in step (5)Scheme GtimeIn all nodes, the node currently traversed is denoted as vk, k expression node identification.At this point, if described in step (1)Adjacency matrix A [k] [det]=1 can assert node v then when f (det)=f (k) described in step (2)detMeetTenant's security isolation.
Further, according to the active detecting method of the multi-tenant cloud platform security isolation, it is characterised in that: step(42), user group security isolation described in (52) detects, and is according to the detection based on user group security isolation described in step (2)Benchmark, the node v that queue Detect head end to be detected described in detecting step (42), (52) is readdetWhether user group is metSecurity isolation.
For node vdetSpecific detection process include: to scheme G described in traversal step (3)initOr described in step (5)Scheme GtimeIn all nodes, the node currently traversed is denoted as vk, k expression node identification.At this point, if described in step (1)Adjacency matrix A [k] [det]=1, then only y (k)=y (det) > 0 or y (k)=1 and y (det) described in step (2)!=1 ory(k)!When=0 and y (det)=0, node v can be assertdetMeet user group security isolation.
The present invention takes above technical scheme, has the following characteristics that
(1) this method carries out organization modeling to the configuration status of multi-tenant cloud platform in graph form, and establishes and be based onThe detection benchmark of tenant's security isolation and the detection benchmark based on user group security isolation, then isolation detection with high safety, withFind that potential isolated failure threatens the harm caused by tenant in cloud in time.
(2) the main dynamic measuring method of the security isolation inspection proposed through the invention, can detect that in the cloud environment of dynamic changePotential isolated failure that may be present threatens, to make up the deficiency of existing method, improves the safe energy of multi-tenant cloud platformPower.
Detailed description of the invention
Fig. 1 shows a kind of active detecting method flow diagram of multi-tenant cloud platform security isolation provided by the invention;
Fig. 2 shows system platform configuration diagrams provided in an embodiment of the present invention;
Fig. 3 shows a kind of implementation of the active detecting method of multi-tenant cloud platform security isolation provided in an embodiment of the present inventionFrame diagram.
Specific embodiment
In order to which target of the invention, technical solution and advantage is more clearly understood, below with reference to the embodiment of the present inventionIn attached drawing the present invention will be described in further detail.It should be appreciated that specific embodiment described herein is only used toIt explains the present invention, is not intended to limit the present invention.Based on the embodiments of the present invention, those skilled in the art are not making woundAll other embodiment obtained under the premise of the property made labour, shall fall within the protection scope of the present invention.
Currently, with the fast development and change of IT infrastructure the relevant technologies, the appearance based on lightweight virtualization technologyDevice engine Docker is used by the more and more companies of every profession and trade.Container (Container) is a kind of New Virtual technology, is madeThe mechanism such as the namespace and cgroup that are supported with linux kernel itself realize the isolation of environment and resource, have flexibilityFeature high, deployment is convenient.It, can be based on appearance by integrating the distributed container cluster Managed Solution using kubernetes as representativeDevice technology building is to provide service as the PaaS cloud of target.
The present embodiment is based on container engine docker and builds cloud platform, manages container cluster, calico by kubernetesThree-layer network project management container cluster network, the NameSpace label and Network provided in conjunction with kubernetes platformPolicy mechanism, realizes the tenant network isolation requirement of cloud platform, and disposing consistency warehouse etcd cooperates with each node of platformWork.
Based on above-mentioned implementation environment, the isolation detection method proposed is applied to container cloud platform by the embodiment of the present invention,And security isolation detection is carried out for tenant network.In the present embodiment, if all users are assigned in management granularity in cloudIn dry group, every group is known as a tenant (tenant), and tenant may include user group again.Therefore, user can be divided into three in cloudClass: including cloud platform administrator, tenant administrator and ordinary user.
Fig. 2 is the platform architecture schematic diagram of the embodiment of the present invention.The present embodiment includes 5 steps.
(1) graph model is defined.Under the PAAS cloud environment of kubernetes cluster building, dummy node is referred to as POD,Therefore using POD be used as the fundamental node unit of cloud platform management, then building cloud platform in each user node state it is orientedGraph model G:
G=(V, E, C)
V indicates the cluster of the node of be provided with service in cloud platform, and any node meets v in clusteri∈V;E indicates eachSide between node, there are any bar side e in clusteri,jIndicate node viIt may have access to another node vj, ei,j∈E;C indicates sectionConnected relation between point, ci,jIndicate node viWith vjBetween connected relation, ci,j∈C.I, j indicate the mark of two nodes.
When constructing Directed Graph Model, if ci,j=1, then it is assumed that v in figure GiWith vjBetween there are side ei,jIf ci,j=0, then viWith viBetween be not present side ei,j.Above-mentioned Directed Graph Model G is stored in the form of two-dimentional adjacency matrix, is denoted as:
(2) definition detection benchmark: in multi-tenant cloud platform, there are multiple tenants to belong to different rents by Admin AdministrationUnauthorized access can not be initiated between the node at family;Inside each tenant, there are multiple user groups, by administrator or tenant administratorManagement, the node for belonging to different user groups can not initiate unauthorized access.Therefore, security isolation includes two class scenes: first is that differentSecurity isolation between tenant need to verify the potential threat that whether there is unauthorized access between each tenant;Second is that same tenantSecurity isolation between interior different user groups need to be verified and whether there is the latent of unauthorized access in same tenant between different user groupsIt is threatening.For the above scene, detection benchmark based on tenant's security isolation is defined respectively and based on user group security isolationDetect benchmark.
(21) based on the detection benchmark of tenant's security isolation: when attacker obtains in administrator right or multi-tenant cloud platformThere are when malice administrator, the configuration of multi-tenant cloud platform can be maliciously tampered in portion, make to originally belong to different tenants node itBetween can initiate unauthorized access, tenant's security isolation is destroyed.Therefore, the correct mapping relationship f for establishing tenant and node, asThe detection benchmark of tenant's security isolation.If n indicates the mark of node, tenentnIndicate the mark for being identified as the node tenant of nKnow, mapping relations be denoted as:
F (n)=tenentn
(22) based on the detection benchmark of user group security isolation: for the user for belonging to same tenant in multi-tenant cloud platformGroup, definition L are user group grade, and L value can be uppr, normal or lower.L value may have access to non-for the user group of upprAll user groups of uppr, can not be by other users group access;L value be normal user group can be all with L value normal andMark value is all that the user group of x is exchanged visits, and x is the integer greater than zero;L value be lower user group can by all user's group access,It can not initiate to access.For the node in user group, if n indicates the mark of node, levelnIndicate the mark for being identified as the node of nNote value, then the mark value level of node nnIs defined as:
When attacker obtains inside administrator or tenant's administrator right and multi-tenant cloud platform, there are malice administratorsOr when malice tenant administrator, the grade configuration of user group can be maliciously tampered, between the node for making different grades of user groupUnauthorized access can be initiated, user group security isolation is destroyed.Therefore, the correct of each user group and node in same tenant is establishedMapping relations are denoted as by mapping relations y as the detection benchmark of user group security isolation:
Y (n)=leveln
(3) original state generates: acquiring the configuration data of all nodes in multi-tenant cloud platform under original state.According to stepSuddenly Directed Graph Model G=(V, E, C) defined in (1) traverses configuration data and records current traverse node viWith other sectionsPoint vjConnected relation ci,j∈C.If viWith vjIt is connected to, then ci,j=1, otherwise ci,j=0, then with step (1) the definition structureBuild the graph model G based on original stateinit
In the present embodiment, data source is each node in container cloud cluster, and the specific data that acquire include: nodeNameSpace label;The Role label of node;The corresponding isolation strategy of node R ole label;Time corresponding to current network stateStamp.
Finally obtained data set such as table 1 shows, gives multi-tenant cloud platform node configuration data sample.
1 multi-tenant cloud platform node configuration data sample of table
The configuration of Namespace attribute provides interface by Namespace isolation mech isolation test in kubernetes platform in table 1,The setting of Role attribute provides interface by Network Policy mechanism in kubernetes platform.
Oriented state diagram G can then be constructedinit, and obtain adjacency matrix:
(4) benchmark, digraph G described in detecting step (3) initial isolation detection: are detected described in foundation step (2)initWhether security isolation is met, and detection process includes initialization vertex ticks and initialization two stages of nodal test:
(41) vertex ticks process is initialized are as follows: successively traversal step (3) the digraph GinitIn all nodes.TimeIt is node to be detected by current accessed vertex ticks, and be inserted into queue Detect to be detected from tail end during going through.
Detect={ pod1, pod2, pod3 }
(42) nodal test process is initialized are as follows: be successively read node up to queue is from queue Detect head end to be detectedThe current node that reads is denoted as v by skydet.Detect vdetWhether security isolation demand is met, and upon completion of the assays by vdetFrom the beginningEnd removes queue Detect to be detected.Specific detection content includes: the tenant's security isolation inspection executed based on detection benchmark (21)It surveys, the user group security isolation detection executed based on detection benchmark (22).
The detection of tenant's security isolation is first carried out: traversal step schemes G described in (3)initIn all nodes, by current timeThe node gone through is denoted as vk, k expression node identification.At this point, if adjacency matrix A [k] [det]=1 described in step (1), whenAnd if only if can assert node v when f (det)=f (k) described in step (2)detMeet tenant's security isolation;
Then execute the state-detection based on region security domain: traversal step schemes G described in (3)initIn all sectionsThe node currently traversed is denoted as v by pointk, k expression node identification.At this point, if adjacency matrix A [k] described in step (1)[det]=1, then only y (k)=y (det) > 0 or y (k)=1 and y (det) described in step (2)!=1 or y (k)!=0 and y(det)=0 when, it can assert node vdetMeet user group security isolation.
(5) isolation detection when running: being denoted as time for the current time of running, if at the time moment, administrator or tenant's pipeReason person, which implements configuration operation to cloud platform, causes digraph state that will update, then current digraph is denoted as Gtime, and according toBenchmark is detected according to described in step (2), detects G when updatingtimeWhether security isolation demand is met, and detection process includes updatingWhen vertex ticks and update when two stages of nodal test:
(51) update when node labeling process are as follows: described in step (5) configuration operation include creation example, delete example,It creates security strategy, delete security strategy.When in cloud platform administrator or tenant administrator implement aforesaid operations when, increment is moreNew graph model Gtime.The incremental update, when implementation, specifically may include following four classes situation:
Side insertion: make step (5) the graph model G when creating or deleting new security strategytimeIn node viIt may have access tovjWhen, connected relation c described in setting steps (1)i,j=1, make adjacency matrix A [i] [j]=1, thereby executing side insertion.
Edge contract: make step (5) the graph model G when creating or deleting new security strategytimeIn node viIt can not visitAsk vjWhen, connected relation c described in setting steps (1)i,j=0, make adjacency matrix A [i] [j]=0, thereby executing edge contract.
Point insertion: when creating new example, in step (5) the graph model GtimeMiddle insertion new node, is denoted as vnew, newIndicate node identification.If remembering GtimeIn other nodes be identified as k, adjacency matrix A [new] [k] described in initialization step (1)=0 and A [new] [k]=0.As node vnewIt may have access to vkWhen, connected relation c described in setting steps (1)new,k=1, make to abutMatrix A [new] [k]=1;As node vkIt may have access to vnewWhen, connected relation c is setk,new=1, make adjacency matrix A [k] [new]=1.Thereby executing an insertion.
Point deletion: when deleting new example, in step (5) the graph model GtimeThe corresponding node of middle positioning example,It is denoted as vdel, del expression node identification.At this time by the way that del=-1 is arranged by node vdelLabeled as failure, it is not involved in detection.FromAnd execute point deletion.
For example, if tenant administrator modifies pod3 in table 1 and correspond to Ingress attribute as " { } ", G at this timetimeIt is corresponding adjacentMatrix is connect to be expressed as:
Then, node will be updated and be inserted into queue Detect to be detected from tail end.At this point, then being marked since pod3 is updatedIt is denoted as node to be detected, Detect queue is added.
Detect={ pod3 }
(52) node detection process when updating are as follows: be successively read node up to queue is from queue Detect head end to be detectedThe current node that reads is denoted as v by skydet.Detect vdetWhether security isolation demand is met, and upon completion of the assays by vdetFrom the beginningEnd removes queue Detect to be detected.Specific detection content includes: the tenant's security isolation inspection executed based on detection benchmark (21)It surveys, the user group security isolation detection executed based on detection benchmark (22).
The detection of tenant's security isolation is first carried out: traversal step schemes G described in (3)timeIn all nodes, by current timeThe node gone through is denoted as vk, k expression node identification.At this point, if adjacency matrix A [k] [det]=1 described in step (1), whenAnd if only if can assert node v when f (det)=f (k) described in step (2)detMeet tenant's security isolation;
Then execute the state-detection based on region security domain: traversal step schemes G described in (3)timeIn all sectionsThe node currently traversed is denoted as v by pointk, k expression node identification.At this point, if adjacency matrix A [k] described in step (1)[det]=1, then only y (k)=y (det) > 0 or y (k)=1 and y (det) described in step (2)!=1 or y (k)!=0 and y(det)=0 when, it can assert node vdetMeet user group security isolation.
Above embodiments are provided just for the sake of the description purpose of the present invention, and are not intended to limit the scope of the invention.ThisThe range of invention is defined by the following claims.It does not depart from spirit and principles of the present invention and the various equivalent replacements made and repairsChange, should all cover within the scope of the present invention.

Claims (4)

Translated fromChinese
1.一种多租户云平台安全隔离的主动检测方法,其特征在于:包括以下步骤:1. an active detection method for multi-tenant cloud platform security isolation, is characterized in that: comprise the following steps:(1)定义图模型:定义包含多租户云平台中各个节点的有向图模型G=(V,E,C),V表示云平台中所有提供服务的节点的集群,集群中任一节点满足vi∈V;E表示各节点之间的边,集群中存在任一条边ei,j即表示节点vi可访问另一节点vj,ei,j∈E;C表示节点之间的连通关系,ci,j表示节点vi与vj之间的连通关系,ci,j∈C,i,j表示两个节点的标识;(1) Define a graph model: define a directed graph model G=(V, E, C) that includes each node in the multi-tenant cloud platform, where V represents the cluster of all nodes that provide services in the cloud platform, and any node in the cluster satisfies the vi ∈V; E represents the edge between nodes, and the existence of any edge ei,j in the cluster means that node vi can access another node vj , ei,j ∈ E; C represents the connection between nodes Connected relationship, ci,j represents the connected relationship between nodes vi and vj , ci,j ∈ C, i,j represents the identity of the two nodes;构建有向图模型时,若ci,j=1,则认为图G中vi与vj之间存在边ei,j,若ci,j=0,则vi与vj之间不存在边ei,j;上述有向图模型G以二维邻接矩阵的形式存储,记为:When constructing a directed graph model, if ci,j =1, it is considered that there is an edge ei, j between vi and vj in the graph G; if ci,j =0, then there is an edge e i,j between vi and vj There is no edge ei,j ; the above directed graph model G is stored in the form of a two-dimensional adjacency matrix, denoted as:(2)定义检测基准:在多租户云平台中,存在多个租户,由管理员管理,属于不同租户的节点之间不可发起非法访问;每个租户内部,存在多个用户组,由管理员或租户管理员管理,属于不同用户组的节点不可发起非法访问;安全隔离包含两类场景:一是不同租户之间的安全隔离,需验证各个租户之间是否存在非法访问的潜在威胁;二是同一租户内不同用户组之间的安全隔离,需验证同一租户内不同用户组之间是否存在非法访问的潜在威胁;针对以上场景,分别定义基于租户安全隔离的检测基准和基于用户组安全隔离的检测基准;(2) Define the detection benchmark: In the multi-tenant cloud platform, there are multiple tenants, which are managed by the administrator, and illegal access cannot be initiated between nodes belonging to different tenants; within each tenant, there are multiple user groups, which are managed by the administrator Or managed by tenant administrators, nodes belonging to different user groups cannot initiate illegal access; security isolation includes two scenarios: one is the security isolation between different tenants, and it is necessary to verify whether there is a potential threat of illegal access between each tenant; For security isolation between different user groups in the same tenant, it is necessary to verify whether there is a potential threat of illegal access between different user groups in the same tenant. For the above scenarios, the detection benchmark based on tenant security isolation and the security isolation based on user group are defined respectively. testing benchmark;(21)基于租户安全隔离的检测基准:当攻击者获得管理员权限或多租户云平台内部存在恶意管理员时,多租户云平台的配置可被恶意篡改,使原本属于不同租户的节点之间可以发起非法访问,租户安全隔离被破坏,建立租户与节点的正确映射关系f,作为租户安全隔离的检测基准;若n表示节点的标识,tenentn表示标识为n的节点所属租户的标识,将映射关系记为:(21) Detection benchmark based on tenant security isolation: When an attacker obtains administrator privileges or there is a malicious administrator inside the multi-tenant cloud platform, the configuration of the multi-tenant cloud platform can be maliciously tampered with, causing nodes that originally belong to different tenants to be tampered with. Illegal access can be initiated, tenant security isolation is destroyed, and the correct mapping relationship f between tenants and nodes is established, as the detection benchmark for tenant security isolation; if n represents the identity of the node, tenantn represents the identity of the tenant to which the node with identity n belongs. The mapping relationship is recorded as:f(n)=tenentnf(n)=tenentn ;(22)基于用户组安全隔离的检测基准:针对多租户云平台中属于同一租户的用户组,定义L为用户组等级,L值为uppr、normal或lower;L值为uppr的用户组可访问非uppr的所有用户组,不可被其他用户组访问;L值为normal的用户组可与L值同为normal且标记值同为x的用户组互访,x为大于零的整数;L值为lower的用户组可被所有用户组访问,不可发起访问;对于用户组内的节点,若n表示节点的标识,leveln表示标识为n的节点的标记值,则节点n的标记值leveln定义为:(22) Detection benchmark based on user group security isolation: For user groups belonging to the same tenant in the multi-tenant cloud platform, define L as the user group level, L value is uppr, normal or lower; user groups whose L value is uppr can access All user groups that are not uppr cannot be accessed by other user groups; a user group whose L value is normal can visit each other with a user group whose L value is both normal and whose tag value is the same as x, where x is an integer greater than zero; L value is Lower user groups can be accessed by all user groups and cannot initiate access; for nodes in a user group, if n represents the node's identifier, and leveln represents the tag value of the node with the tag n, then the tag value of node n is defined as leveln . for:当攻击者获得管理员或租户管理员权限,以及多租户云平台内部存在恶意管理员或恶意租户管理员时,用户组的等级配置可被恶意篡改,使不同等级的用户组的节点之间可发起非法访问,用户组安全隔离被破坏;建立同一租户内各个用户组与节点的正确映射关系y,作为用户组安全隔离的检测基准,将映射关系记为:When an attacker obtains administrator or tenant administrator rights, and there are malicious administrators or malicious tenant administrators in the multi-tenant cloud platform, the level configuration of user groups can be maliciously tampered with, so that nodes of different levels of user groups can be tampered with. When illegal access is initiated, the security isolation of user groups is destroyed; the correct mapping relationship y between each user group and node in the same tenant is established, as the detection benchmark for security isolation of user groups, and the mapping relationship is recorded as:y(n)=levelny(n)=leveln ;(3)初始状态生成:采集初始状态下多租户云平台中所有节点的配置数据,依据步骤(1)中所定义的有向图模型G=(V,E,C),遍历配置数据并记录当前遍历节点vi与其他节点vj的连通关系ci,j∈C;若vi与vj连通,则ci,j=1,否则ci,j=0,然后以步骤(1)所述定义构建基于初始状态的图模型Ginit(3) Initial state generation: collect the configuration data of all nodes in the multi-tenant cloud platform in the initial state, and traverse the configuration data and record it according to the directed graph model G=(V, E, C) defined in step (1). The connectivity relationship ci,jCbetween the current traversal node viand other nodes vj is The definition constructs a graph model Ginit based on the initial state;(4)初始隔离检测:依据步骤(2)中所述检测基准,检测步骤(3)中所述有向图Ginit是否满足安全隔离,检测过程包括初始化节点标记和初始化节点检测两个阶段:(4) Initial isolation detection: According to the detection benchmark described in step (2), to detect whether the directed graph Ginit described in step (3) satisfies security isolation, the detection process includes two stages of initializing node marking and initializing node detection:(41)初始化节点标记过程为:依次遍历步骤(3)所述有向图Ginit中的所有节点;遍历过程中,将当前访问节点标记为待检测节点,并从尾端插入待检测队列Detect;(41) The initialization node marking process is: traverse all the nodes in the directed graph Ginit described in step (3) in turn; in the traversing process, mark the current access node as the node to be detected, and insert the queue Detect to be detected from the end ;(42)初始化节点检测过程为:从待检测队列Detect头端依次读取节点直至队列为空,将当前读取节点记为vdet,检测vdet是否符合安全隔离需求,并在检测完成后将vdet从头端移出待检测队列Detect,具体检测内容包括:基于检测基准(21)执行的租户安全隔离检测,基于检测基准(22)执行的用户组安全隔离检测;(42) The initial node detection process is: read the nodes sequentially from the head end of the queue Detect to be detected until the queue is empty, record the current read node as vdet , check whether vdet meets the security isolation requirements, and after the detection is completed, vdet removes the queue to be detected Detect from the headend, and the specific detection contents include: tenant security isolation detection performed based on the detection benchmark (21), and user group security isolation detection performed based on the detection benchmark (22);(5)运行时隔离检测:将当前运行时刻记为time,若在time时刻,管理员或租户管理员对云平台实施配置操作导致有向图状态将发生更新,则将当前有向图记为Gtime,并依据步骤(2)中所述检测基准,在更新时检测Gtime是否满足安全隔离需求,检测过程包括更新时节点标记和更新时节点检测两个阶段:(5) Run-time isolation detection: record the current running time as time. If the administrator or tenant administrator performs configuration operations on the cloud platform at time, the state of the directed graph will be updated, and the current directed graph is recorded as Gtime , and according to the detection benchmark described in step (2), check whether Gtime meets the security isolation requirement during update, and the detection process includes two stages: node marking during update and node detection during update:(51)更新时节点标记过程为:步骤(5)中所述配置操作包括创建实例、删除实例、创建安全策略、删除安全策略;当云平台中的管理员或租户管理员实施上述操作时,增量更新图模型Gtime,并将被更新节点标记为待检测节点,从尾端插入待检测队列Eetect;(51) The node marking process during the update is: the configuration operations described in step (5) include creating an instance, deleting an instance, creating a security policy, and deleting a security policy; when the administrator in the cloud platform or the tenant administrator implements the above operations, Incrementally update the graph model Gtime , mark the updated node as the node to be detected, and insert the queue to be detected Eetect from the end;(52)更新时节点检测过程为:从待检测队列Eetect头端依次读取节点直至队列为空,将当前读取节点记为vdet,检测vdet是否符合安全隔离需求,并在检测完成后将vdet从头端移出待检测队列Detect,具体检测内容包括:基于检测基准(21)执行的租户安全隔离检测,基于检测基准(22)执行的用户组安全隔离检测。(52) The node detection process during the update is: read the nodes in sequence from the head end of the queue to be detectedEetect until the queue is empty, record the current read node as vdet , check whether v det meets the security isolation requirements, and when the detection is completed Then, vdet is removed from the queue to be detected Detect from the head end, and the specific detection content includes: tenant security isolation detection based on the detection benchmark (21), and user group security isolation detection based on the detection benchmark (22).2.根据权利要求1所述的多租户云平台安全隔离的主动检测方法,其特征在于:步骤(51)中所述增量更新模型Gtime,实施时具体包含以下四类情况,i,j表示节点的标识:2. the active detection method of multi-tenant cloud platform security isolation according to claim 1, is characterized in that: incremental update model Gtime described in step (51), specifically includes following four kinds of situations during implementation, i, j Represents the ID of the node:边插入:当创建或删除新的安全策略使步骤(5)所述图模型Gtime中的节点vi可访问vj时,设置步骤(1)中所述连通关系ci,j=1,使邻接矩阵A[i][j]=1,从而执行边插入;Edge insertion: when a new security policy is created or deleted to make the node vi in the graph model Gtime in step (5) accessible to vj , set the connectivity relationship ci,j =1 in step (1), Make the adjacency matrix A[i][j]=1, thereby performing edge insertion;边删除:当创建或删除新的安全策略使步骤(5)所述图模型Gtime中的节点vi不可访问vj时,设置步骤(1)中所述连通关系ci,j=0,使邻接矩阵A[i][j]=0,从而执行边删除;Edge deletion: when the creation or deletion of a new security policy makes the node vi in the graph model Gtime in step (5) inaccessible to vj , set the connectivity relationship ci,j =0 in step (1), Make the adjacency matrix A[i][j]=0, thereby performing edge deletion;点插入:当创建新实例时,在步骤(5)所述图模型Gtime中插入新节点,记为vnew,new表示节点标识;若记Gtime中其他节点的标识为k,初始化步骤(1)中所述邻接矩阵A[new][k]=0且A[new][k]=0;当节点vnew可访问vk时,设置步骤(1)中所述连通关系cnew,k=1,使邻接矩阵A[new][k]=1;当节点vk可访问vnew时,设置连通关系ck,new=1,使邻接矩阵A[k][new]=1,从而执行点插入;Point insertion: when a new instance is created, a new node is inserted in the graph model Gtime described in step (5), denoted as vnew , and new represents the node identifier; if the identifier of other nodes in Gtime is recorded as k, the initialization step ( 1) The adjacency matrix A[new][k]=0 and A[new][k]=0; when the node vnew can access vk , set the connectivity relationship cnew in step (1), k =1, make the adjacency matrix A[new ][k]=1; when the nodevk can access vnew, set the connectivity relationshipck, new =1, make the adjacency matrix A[k][new]=1, thereby performing point insertion;点删除:当删除新实例时,在步骤(5)所述图模型Gtime中定位该实例对应的节点,记为vdel,del表示节点标识;此时通过设置del=-1将节点vdel标记为失效,不参与检测;从而执行点删除。Point deletion: when deleting a new instance, locate the node corresponding to the instance in the graph model Gtime described in step (5), denoted as vdel , and del represents the node identifier; at this time, by setting del=-1, the node vdel Mark as invalid and do not participate in detection; thus perform point deletion.3.根据权利要求1所述的多租户云平台安全隔离的主动检测方法,其特征在于:步骤(42)、(52)中所述租户安全隔离检测,是依据步骤(2)中所述基于租户安全隔离的检测基准,检测步骤(42)、(52)中所述待检测队列Eetect头端读取到的节点vdet是否满足租户安全隔离;针对节点vdet的具体检测过程包括:遍历步骤(3)中所述图Ginit或步骤(5)中所述图Gtime中的所有节点,将当前遍历到的节点记为vk,k表示节点标识;此时,若步骤(1)中所述邻接矩阵A[k][det]=1,则当且仅当步骤(2)中所述f(det)=f(k)时,可认定节点vdet满足租户安全隔离。3. The active detection method of multi-tenant cloud platform security isolation according to claim 1, characterized in that: the tenant security isolation detection described in steps (42) and (52) is based on the The detection benchmark of tenant security isolation, whether the node vdet read by the head end of the queue Eetect to be detected in the detection steps (42) and (52) satisfies the tenant security isolation; the specific detection process for the node vdet includes: traversing step For all nodes in the graph Ginit described in (3) or the graph Gtime described in step (5), the currently traversed node is denoted as vk , and k represents the node identifier; at this time, if in step (1) The adjacency matrix A[k][det]=1, then if and only if f(det)=f(k) in step (2), it can be determined that the node vdet satisfies the tenant security isolation.4.根据权利要求1所述的多租户云平台安全隔离的主动检测方法,其特征在于:步骤(42)、(52)中所述用户组安全隔离检测,是依据步骤(2)中所述基于用户组安全隔离的检测基准,检测步骤(42)、(52)中所述待检测队列Detect头端读取到的节点vdet是否满足用户组安全隔离;针对节点vdet的具体检测过程包括:遍历步骤(3)中所述图Ginit或步骤(5)中所述图Gtime中的所有节点,将当前遍历到的节点记为vk,k表示节点标识;此时,若步骤(1)中所述邻接矩阵A[k][det]=1,则仅当步骤(2)中所述y(k)=y(det)>0或y(k)=1且y(det)!=1或y(k)!=0且y(det)=0时,可认定节点vdet满足用户组安全隔离。4. the active detection method of multi-tenant cloud platform security isolation according to claim 1, is characterized in that: user group security isolation detection described in step (42), (52), is according to step (2) described in Based on the detection benchmark of user group security isolation, check whether the node vdet read by the head end of the queue Detect to be detected in steps (42) and (52) satisfies the user group security isolation; the specific detection process for the node vdet includes: : traverse all nodes in the graph Ginit described in step (3) or graph Gtime described in step (5), and record the currently traversed node as vk , where k represents the node identifier; at this time, if step ( The adjacency matrix A[k][det]=1 in 1), then only if y(k)=y(det)>0 or y(k)=1 and y(det) in step (2) ! =1 or y(k)! =0 and y(det)=0, it can be determined that the node vdet satisfies the user group security isolation.
CN201811473272.8A2018-12-042018-12-04Active detection method for multi-tenant cloud platform security isolationActiveCN109412866B (en)

Priority Applications (1)

Application NumberPriority DateFiling DateTitle
CN201811473272.8ACN109412866B (en)2018-12-042018-12-04Active detection method for multi-tenant cloud platform security isolation

Applications Claiming Priority (1)

Application NumberPriority DateFiling DateTitle
CN201811473272.8ACN109412866B (en)2018-12-042018-12-04Active detection method for multi-tenant cloud platform security isolation

Publications (2)

Publication NumberPublication Date
CN109412866Atrue CN109412866A (en)2019-03-01
CN109412866B CN109412866B (en)2020-07-28

Family

ID=65456980

Family Applications (1)

Application NumberTitlePriority DateFiling Date
CN201811473272.8AActiveCN109412866B (en)2018-12-042018-12-04Active detection method for multi-tenant cloud platform security isolation

Country Status (1)

CountryLink
CN (1)CN109412866B (en)

Citations (6)

* Cited by examiner, † Cited by third party
Publication numberPriority datePublication dateAssigneeTitle
US8839345B2 (en)*2008-03-172014-09-16International Business Machines CorporationMethod for discovering a security policy
CN104092565A (en)*2014-06-242014-10-08复旦大学 A Multitenant Policy-Driven Software-Defined Networking Approach for Cloud Data Centers
CN106569895A (en)*2016-10-242017-04-19华南理工大学Construction method of multi-tenant big data platform based on container
CN107104963A (en)*2017-04-252017-08-29华中科技大学Towards the trusted controller framework and its operating method of cloud environment multi-tenant network
US20170261264A1 (en)*2017-05-252017-09-14Northeastern UniversityFault diagnosis device based on common information and special information of running video information for electric-arc furnace and method thereof
CN107689953A (en)*2017-08-182018-02-13中国科学院信息工程研究所A kind of vessel safety monitoring method and system towards multi-tenant cloud computing

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication numberPriority datePublication dateAssigneeTitle
US8839345B2 (en)*2008-03-172014-09-16International Business Machines CorporationMethod for discovering a security policy
CN104092565A (en)*2014-06-242014-10-08复旦大学 A Multitenant Policy-Driven Software-Defined Networking Approach for Cloud Data Centers
CN106569895A (en)*2016-10-242017-04-19华南理工大学Construction method of multi-tenant big data platform based on container
CN107104963A (en)*2017-04-252017-08-29华中科技大学Towards the trusted controller framework and its operating method of cloud environment multi-tenant network
US20170261264A1 (en)*2017-05-252017-09-14Northeastern UniversityFault diagnosis device based on common information and special information of running video information for electric-arc furnace and method thereof
CN107689953A (en)*2017-08-182018-02-13中国科学院信息工程研究所A kind of vessel safety monitoring method and system towards multi-tenant cloud computing

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
张坤: "《面向多租户应用的云数据隐私保护机制研究》", 《山东大学博士论文》*

Also Published As

Publication numberPublication date
CN109412866B (en)2020-07-28

Similar Documents

PublicationPublication DateTitle
CN115039098B (en) Fuzzy network detection pattern matching
US10956566B2 (en)Multi-point causality tracking in cyber incident reasoning
US11941054B2 (en)Iterative constraint solving in abstract graph matching for cyber incident reasoning
Khan et al.Cloud log forensics: Foundations, state of the art, and future directions
Wang et al.Fog computing: Issues and challenges in security and forensics
US20200120118A1 (en)Endpoint inter-process activity extraction and pattern matching
US11330007B2 (en)Graphical temporal graph pattern editor
CN106687971A (en)Automated code lockdown to reduce attack surface for software
Fan et al.SBBS: A secure blockchain-based scheme for IoT data credibility in fog environment
US12250230B2 (en)Lateral movement analysis using certificate private keys
CN108600163A (en)A kind of cloud environment distributed hash chain framework and cloud data integrity verification method
US10587652B2 (en)Generating false data for suspicious users
US20230247040A1 (en)Techniques for cloud detection and response from cloud logs utilizing a security graph
Mishra et al.Cloud forensics: State-of-the-art and research challenges
CN117318978A (en)Knowledge graph-based network asset attack surface management method, system, device and storage medium
US12430423B2 (en)Automated generation of labeled training data
Li et al.LogKernel: A threat hunting approach based on behaviour provenance graph and graph kernel clustering
Vu-Minh et al.Performance Evaluation of Decentralized Machine Learning based Network-Based Intrusion Detection System for Internet of Things
Chauhan et al.Emphasizing on various security issues in cloud forensic framework
CN109412866A (en)A kind of active detecting method of multi-tenant cloud platform security isolation
CN116800479A (en)Traceability map construction method and system based on container arrangement system
Wang et al.A novel covert channel detection method in cloud based on XSRM and improved event association algorithm
US12335286B1 (en)Compute environment security monitoring using data collected from a sub-kernel space
Luo et al.Modeling, conflict detection, and verification of a new virtualization role‐based access control framework
Chen et al.FedTag: Towards Automated Attack Investigation Using Federated Learning

Legal Events

DateCodeTitleDescription
PB01Publication
PB01Publication
SE01Entry into force of request for substantive examination
SE01Entry into force of request for substantive examination
GR01Patent grant
GR01Patent grant

[8]ページ先頭

©2009-2025 Movatter.jp