Overload control of microservice-based cloud architectures via differential service and differential retry

A differential retry mechanism in microservice-based cloud architectures adjusts retry parameters based on client and service identity to enhance availability and reduce costs by mitigating denial-of-service attacks.

WO2026239390A1PCT designated stage Publication Date: 2026-11-19RAMOT AT TEL AVIV UNIVERSITY LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
PCT/IL2026/050406
Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
Priority Date
2025-05-14
Filing Date
2026-05-14
Publication Date
2026-11-19

AI Technical Summary

Technical Problem

Existing cloud computing systems face challenges in implementing effective denial-of-service protection, particularly in microservice-based architectures, where service function invocations can fail due to network errors, service overload, or attacks like YoYo, leading to increased processing burden and monetary costs.

Method used

Implementing a differential retry mechanism within a service mesh that adjusts retry parameters based on service-invocation retry parameters, including identity of the invoking microservice, client priority, and denial-of-service risk score, to manage service function invocations and mitigate denial-of-service attacks.

Benefits of technology

Enhances service availability and reduces response time for high-priority applications by selectively retrying service invocations, thereby increasing resistance to denial-of-service attacks and minimizing monetary impact.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure IL2026050406_19112026_PF_FP_ABST
    Figure IL2026050406_19112026_PF_FP_ABST
Patent Text Reader

Abstract

A system of cloud-based interservice invocation, the system comprising a processing circuitry' (PC) configured to: invoke the service function provided by a first cloud sendee; send responsive to a failure of the invoking of the sendee function, perform at least one of: scheduling a reinvocation of the sendee function based on the one or more service invocation retry parameters associated with the invoking, conditionally reinvoking the service function, based on the one or more service invocation retry parameters associated with the invoking.
Need to check novelty before this filing date? Find Prior Art

Description

[0001] OVERLOAD CONTROL OF MICROSERVICE-BASED CLOUD ARCHITECTURES VIA DIFFERENTIAL SERVICE AND DIFFERENTIAL

[0002] RETRY

[0003] TECHNICAL FIELD

[0004] The presently disclosed subject matter relates to use of security of microservices in cloud computing, and in particular to implementation of denial-of-service protection in such systems.

[0005] BACKGROUND

[0006] Problems of secure implementation of cloud computing systems have been recognized in the conventional art and various techniques have been developed to provide solutions.

[0007] GENERAL DESCRIPTION

[0008] According to one aspect of the presently disclosed subject matter there is provided a sy stem of cloud-based interservice invocation, the sy stem comprising a processing circuitry configured to:

[0009] a) invoke a service function provided by a first cloud service; and b) responsive to a failure of the invoking of the service function, perform at least one of:

[0010] i. scheduling a reinvocation of the service function based on one or more service-invocation retry parameters associated with the invoking; or ii, conditionally reinvoking the service function based on the one or more service-invocation retry parameters associated with the invoking.

[0011] In addition to the above features, the system according to this aspect of the presently disclosed subject matter can further comprise one or more of features (i)to (v) listed below, in any desired combination or permutation which is technically possible:

[0012] (i) steps a)–b) being performed by a service mesh, and the processing circuitry being further configured to, prior to step a), obtain a request of invocation, by a second cloud service, of the service function provided by the first cloud sendee;

[0013] (ii) at least one of the one or more service-invocation retry parameters being derivative of at least one of an identity of an invoking microservice, data determined by an invoking microservice or predecessor invoking microservice, and data derivative of the identity of the client system;

[0014] (iii) the data determined by an invoking microservice or predecessor invoking microservice comprising a client-priority ranking or denial-of-service risk score;

[0015] (iv) the data derivative of the identity of the client system being based on network monitoring;

[0016] (v) the service-invocation retry parameters comprising at least one of: a between-retry delay value, a maximum retry count, a probability of retrying, and a maximum value of a randomized between-retry7delay7alue.

[0017] According to another aspect of the presently disclosed subject matter there is provided a processing-circuitry-based method of cloud-based interservice invocation, the method comprising:

[0018] a) invoking a service function provided by a first cloud service; and b) responsive to a failure of the invoking of the service function, performing at least one of:

[0019] i. scheduling a reinvocation of the service function based on one or more service-invocation retry parameters associated with the invoking; or ii. conditionally reinvoking the service function based on the one or more service-invocation retry parameters associated with the invoking.

[0020] This aspect of the disclosed subject matter can further optionally comprise one or more of features (i) to (v) listed above with respect to the system, mutatis mutandis, in any desired combination or permutation which is technically possible.

[0021] According to another aspect of the presently disclosed subject matter there is provided a computer program product comprising a computer-readable non-transitory storage medium containing program instructions which, when read by processing circuitry, cause the processing circuitry to perform a method of cloud¬ based interservice invocation, the method comprising:

[0022] a) invoking a service function provided by a first cloud service; and b) responsive to a failure of the invoking of the service function, performing at least one of:

[0023] i. scheduling a reinvocation of the service function based on one or more service-invocation retry’ parameters associated with the invoking; or ii. conditionally reinvoking the sendee function based on the one or more service-invocation retry parameters associated with the invoking.

[0024] This aspect of the disclosed subject matter can further optionally comprise one or more of features (i) to (v) listed above with respect to the system, mutatis mutandis, in any desired combination or permutation which is technically possible.

[0025] According to one aspect of the presently disclosed subject matter there is provided a system of cloud-based services enabling enhanced load tolerance, the system comprising a processing circuitry configured to:

[0026] a) receive, by a first cloud service, data originated by a client;

[0027] b) determine, based on the data originated by the client, data indicative of service-invocation retry parameters;

[0028] c) invoke a service function provided by a second microservice; and d) responsive to a failure of the invoking of the service function, perform retry of the invoking, based on the service-invocation retry parameters.

[0029] In addition to the above features, the sy stem according to this aspect of the presently disclosed subject matter can further comprise one or more of features (i) to (iii) listed below, in any desired combination or permutation which is technically possible:

[0030] (i) identifying cookie data in the data originated by the client, and utilizing, at least, the identified cookie data in determining the service-invocation retry parameters;

[0031] (ii) determining a denial-of-service risk score based on data originated by the client, and utilizing, at least, the denial-of-service risk score in determining the service-invocation retry parameters;

[0032] (iii) determining a client-priority ranking based on data originated by the client, and utilizing, at least, the client-priority ranking in determining the service-invocation retry parameters.

[0033] According to another aspect of the presently disclosed subject matter there is provided a system of cloud-based services enabling prioritization, the system comprising a processing circuitry' configured to:

[0034] a) receive, by a first cloud service, a first request originated by a first client and a second request originated by a second client;

[0035] b) determine, based on the first request and the second request, data indicative of respective application priority; and

[0036] c) based on the respective application priority, prioritize one of the requests over the other request.

[0037] This aspect of the disclosed subject matter can further optionally comprise one or more of features (i) to (iii) listed above with respect to the system of cloud-based services enabling enhanced load tolerance, mutatis mutandis, in any desired combination or permutation which is technically possible.

[0038] According to another aspect of the presently disclosed subject matter there is provided a processing-circuitry-based method of enabling enhanced load tolerance in cloud-based services, the method comprising:

[0039] a) receiving, by a first cloud service, data originated by a client;

[0040] b) determining, based on the data originated by the client, data indicative of service-invocation retry parameters;

[0041] c) invoking a service function provided by a second microservice; and d) responsive to a failure of the invoking of the sendee function, performing retry of the invoking, based on the service-invocation retry parameters.

[0042] This aspect of the disclosed subject matter can further optionally comprise one or more of features (i) to (iii) listed above with respect to the system of cloud¬ based services enabling enhanced load tolerance, mutatis mutandis, in any desired combination or permutation which is technically possible.

[0043] According to another aspect of the presently disclosed subject matter there is provided a computer program product comprising a computer-readable non-transitory storage medium containing program instructions which, when read by processing circuitry, cause the processing circuitry to perform a method of enabling enhanced load tolerance in cloud-based services, the method comprising:

[0044] a) receiving, by a first cloud service, data originated by a client;

[0045] b) determining, based on the data originated by the client, data indicative of service-invocation retry parameters;

[0046] c) invoking a service function provided by a second microservice; and

[0047] d) responsive to a failure of the invoking of the service function, performing retry of the invoking, based on the service-invocation retry parameters.

[0048] This aspect of the disclosed subject matter can further optionally comprise one or more of features (i) to (iii) listed above with respect to the system of cloud-based services enabling enhanced load tolerance, mutatis mutandis, in any desired combination or permutation which is technically possible.

[0049] BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to understand the invention and to see how it can be carried out in practice, embodiments will be described, by way of non -limiting examples, with reference to the accompanying drawings, in which:

[0051] Fig. 1A illustrates a logical block diagram of an example cloud provider service mesh with differential retry, in accordance with some embodiments of the presently disclosed subject matter;

[0052] Fig. 1B illustrates a block diagram of an additional example cloud provider service mesh with differential retry', in accordance with some embodiments of the presently disclosed subject matter;

[0053] Fig. 1C illustrates a block diagram of an additional example cloud provider sendee mesh with differential retry, in accordance with some embodiments of the presently disclosed subject matter;

[0054] Fig. ID illustrates an example deployment of an internet-based client accessing a microservices-based cloud application, in accordance with some embodiments of the presently disclosed subject matter;

[0055] Fig. 2 illustrates a flow diagram of an example method of differential retry after service invocation failure, in accordance with some embodiments of the presently’ disclosed subject matter;

[0056] Fig. 3 illustrates a logical block diagram of an example application, together with its utilized cloud-based microservice, with enablement of enhanced load tolerance, in accordance with some embodiments of the presently disclosed subject matter; and

[0057] Fig. 4 illustrates a flow’ diagram of an example method of prioritization of cloud¬ based service utilization based on information derived from client data, in accordance with some embodiments of the presently disclosed subject matter.

[0058] DETAILED DESCRIPTION

[0059] In the following detailed description, numerous specific details are set forth in order to provide a thorough understanding of the invention. However, it will be understood by those skilled in the art that the presently disclosed subject matter may be practiced without these specific details. In other instances, well-known methods, procedures, components and circuits have not been described in detail so as not to obscure the presently disclosed subject matter.

[0060] Unless specifically stated otherwise, as apparent from the following discussions, it is appreciated that throughout the specification discussions utilizing terms such as "processing", "computing", "comparing", “invoking”, "determining", "calculating", “receiving”, “providing”, “obtaining”, “emulating” or the like, refer to the action(s) and / or process(es) of a computer that manipulate and / or transform data into other data, said data represented as physical, such as electronic, quantities and / or said data representing the physical objects. The term “computer” should be expansively construed to cover any kind of hardware-based electronic device with data processing capabilities including, by way of non-limiting example, the processor, mitigation unit, and inspection unit therein disclosed in the present application.

[0061] The terms “non -transitory memory" and “non -transitory storage medium” used herein should be expansively construed to cover any volatile or non-volatile computer memory suitable to the presently disclosed subject matter.

[0062] The operations in accordance with the teachings herein may be performed by a computer specially constructed for the desired purposes or by a general-purpose computer specially configured for the desired purpose by a computer program stored in a non-transitory computer-readable storage medium.

[0063] Embodiments of the presently disclosed subject matter are not described with reference to any particular programming language. It will be appreciated that a variety of programming languages may be used to implement the teachings of the presently disclosed subject matter as described herein.

[0064] It is to be understood that the invention is not limited in its application to the details set forth in the description contained herein or illustrated in the drawings. The invention is capable of other embodiments and of being practiced and carried out in various ways. Hence, it is to be understood that the phraseology and terminology employed herein are for the purpose of description and should not be regarded as limiting. As such, those skilled in the art will appreciate that the conception upon which this disclosure is based may readily be utilized as a basis for designing other structures, methods, and systems for carrying out the several purposes of the presently disclosed subject matter.

[0065] It will also be understood that the system according to the invention may be, at least partly, implemented on a suitably programmed computer. Likewise, the invention contemplates a computer program being readable by a computer for executing the method of the invention. The invention further contemplates a non-transitory computer-readable memory tangibly embodying a program of instructions executable by the computer for executing the method of the invention.

[0066] Cloud applications can utilize a microservices-based architecture i.e. a large application can be composed of a plurality of smaller loosely-coupled sendees. In such anarchitecture, each sendee can be developed, deployed and managed independently, and can be responsible for specific tasks. Such a microservices architecture can be present in both public and private cloud architectures.

[0067] The term “service mesh” is herein interpreted to include an infrastructure layer that manages service-to-service communication within a microservices architecture. A service mesh can provide capabilities such as load balancing, service discovery, encryption, observability, policy enforcement etc.

[0068] Microservices can - as needed - establish connections and / or make requests to / invoke other microservices (for example: via the cloud provider’s sendee mesh). The term “service function” is herein interpreted to include an invocable entry’ point provided by a particular microservice. On occasion, upon a first microservice invoking a sendee function provided by’ a second microsendee, the invocation may fail.

[0069] Microservice sendee function invocation failures can result from - for example - network errors (e.g. when a component loses connectivity for a short period of time), an error internal to the service (e.g., during automatic recovery from a crash), or when a service is overloaded and needs to limit incoming traffic.

[0070] Some embodiments of the presently disclosed subject matter enable differential retry mechanisms in microservices-based applications. In the event of a failure to invoke a microsendee, a retry mechanism (located, for example, within the service mesh) can e.g. continue invocation attempts until either the invocation succeeds or until retry policy dictates that retry must halt.

[0071] In this manner, service availability’ can be improved, and suitably low response time can be achieved for high priority applications.

[0072] In some embodiments of the presently disclosed subject matter, differential retry’ mechanisms can include e.g. under which conditions to retry’, or how to schedule a retry, as will be described in detail below.

[0073] “YoYo” is a denial-of-service method characterized by short and periodic high-rate traffic bursts, which can be hard to detect and mitigate due to the reaction time of conventional DDoS mitigation solutions.

[0074] During a YoYo or similar attack, a microservice retry mechanism can be required to manage significantly’ higher traffic peaks - compared to typical traffic fluctuations. Due to pricing structure of cloud services, the processing burden resulting from application of the retry mechanism during attacks can resulting in significant monetary impact.

[0075] In some embodiments of the presently disclosed subject matter, a differential retry mechanism can reduce or block retries of requests initiated by hostile or potentially hostile client entities, thereby resulting in increased resistance to denial-of-service attacks.

[0076] Attention is directed to Fig. 1, which illustrates a logical block diagram of an example cloud provider service mesh with differential retry, in accordance with some embodiments of the presently disclosed subject matter.

[0077] Processing circuitiy (cloud infrastructure) 100 can include a processor 105 and a memory 110.

[0078] Processor 105 can be a suitable hardware-based electronic device with data processing capabilities, such as, for example, a general purpose processor, digital signal processor (DSP), a specialized Application Specific Integrated Circuit (ASIC), one or more cores in a multicore processor, etc. Processor 105 can also consist, for example, of multiple processors, multiple ASICs, virtual processors, combinations thereof etc.

[0079] Memory 110 can be, for example, a suitable kind of volatile and / or non-volatile storage, and can include, for example, a single physical memory component or a plurality of physical memory' components. Memory' 110 can also include virtual memory'. Memory' 110 can be configured to, for example, store various data used in computation.

[0080] Processing circuitry (cloud infrastructure) 100 can be configured to execute several functional modules in accordance with computer-readable instructions implemented on anon-transitory’ computer-readable storage medium. Such functionalmodules are referred to hereinafter as comprised in the processing circuitry. These modules can include, for example, differential retry unit USA, Service 1 120 and Service_2 125.

[0081] Service_1 120 and Service_2 125, can be, for example, microservices executing in a public or private cloud environment (for example, on a cloud-based server - or alternatively: a serverless cloud environment). It is noted that Service_1 120 andService_2 125 can execute on processing circuitry (cloud infrastructure) 100, or on a different platform (e.g. operably connected to processing circuitry (cloud infrastructure) 100.

[0082] Service_1 120 can seek to execute a service function provided by Service_2125.

[0083] To make use of the service, Service_1 120 can e.g. invoke a function provided by the cloud's service mesh. Invoking the service can be represented by a function call, such as:

[0084] Invoke (Service_2, service_functionQ, retryParameters)

[0085] where Service 2 is the name of the service, service functionQ is the function provided by the service, and retryParameters are parameters that can affect behavior of the retry mechanism, as will be described in detail below, with reference to Fig. 2.

[0086] It is noted that in some embodiments, retryParameters can be made available to the service mesh in a manner that does not involve passing function parameters. For example, the service mesh can maintain tables of retry parameter data to be utilized based on e.g. identity of invoking service, client instance, time-of-day etc. or combinations thereof.

[0087] In some embodiments, a configuration facility can provide such tables of retry parameter data, based on e.g. administrator-provided information. In some such embodiments, a configuration facility can configure retry parameters based on runtime detections such as those described herein. In some such embodiments, a configuration facility can configure retry' parameters based on both administrator-provided information and runtime detections.

[0088] Upon execution of the InvokeQ function, processing circuitry (cloud infrastructure) 100 can e.g. identify the service and service function requested and attempt to connect to the service and invoke the service function (e.g. in a cloud - provider-specific manner).

[0089] Processing circuitry (cloud infrastructure) 100 can include differential retry unit 115, Differential retry unit 115A can control retries of interservice invocations - as will be described in detail below, with reference to Fig. 2.

[0090] It is noted that Fig. 1A presents two services with a single invocation between for simplicity of description, and that actual systems can include multiple services and multiple interservice service functions.

[0091] It is further noted that microservices can include high-load / high-cost functionality such as large-language models (LLMs) or external calls to services, and that - for example - in period of high system load it can be advantageous to limit invocations of such microservices (or of ‘‘‘predecessor” microservices e.g. microservices which invoke high-load / high-cost microservices etc.).

[0092] Fig. 1B illustrates a block diagram of an additional example cloud provider service mesh with differential retry, in accordance with some embodiments of the presently disclosed subject matter.

[0093] In the system of Fig. IB, differential retry unit 115B is part of Service_1 120B rather than the service mesh provided by processing circuitry (cloud infrastructure) 100B. Accordingly, Service_1 120B manages and performs retries. As well, retry parameters can also be received by / maintained by Service_1 120B. Accordingly, the function call can be given as:

[0094] Invoke (Service_2, service_function())

[0095] Fig. 1C illustrates a block diagram of an additional example cloud provider service mesh with differential retry, in accordance with some embodiments of the presently disclosed subject matter.

[0096] In the system of Fig. 1C, differential retry unit 115C is part of Service 1 120B (as in Fig. IB). In this case, the cloud infrastructure provides only an interservice communication facility, which can be represented e.g. by the function calls:

[0097] TransmitData (Service_2)

[0098] ReceiveData (Service_2)

[0099] i.e. for respectively transmitting data to service ___2 and receiving data from sen- ice 2.

[0100] In this case. Service_1 and Service_2 are themselves responsible for conducting / implementing a protocol of intersendee invocation.

[0101] Fig. ID illustrates an example deployment of an internet-based client accessing a microservices-based cloud application, in accordance with some embodiments of the presently disclosed subject matter.

[0102] HTTP-based client 100D can be e.g. a web browser executing on a personal computer or mobile device. HTTP-based client 100D can access a cloud-based application executing on Processing Circuitry (Cloud Infrastructure) 310 (comprising processor 1350 and memory 1300) via the internet 105D.

[0103] The cloud-based application can be - for example - an online shopping website, or any other application. The cloud-based application can be implemented using a number of microservices – in this example: service_1 110D, service_2 115D, service_3 120D, service_4 125D. In this example, service_1 110D can invoke service_2 115D, and service_3 125D; service_2 115D can invoke service_3 125D and service_4 120D; and service_3 125D can invoke service_4 120D. In this example, HTTP-based client 100D can invoke service_1 110D.

[0104] It is noted that a microservice that invokes another microservice is termed an “invoking microsendee ’.

[0105] It is further noted that where a first microservice invokes a second microservice, which then invokes a third microservice, the first microservice is termed a "‘predecessor invoking microsendee”. By way of non-limiting example: if service 1 invokes service_2, which then invokes service_3, which then invokes service_4, then service_1 and

[0106] service 2 are both termed “predecessor invoking microsendees” to the invocation of service 4.

[0107] Each invocation of a sendee function of a particular microsenice can be associated with one or more retry parameters. It is noted that the retry parameters associated with invocation of a particular service function can be derivative of, for example:

[0108] a. an identity of an invoking microservice

[0109] b. data determined by an invoking microservice or predecessor invoking microservice (for example: a client priority or denial-of-service risk, as determined by a microservice based on user transaction data)

[0110] c. data derivative of the identity’ of the client system (for example: as received from a management system, based on network monitoring e.g of network traffic flows and network performance)

[0111] It is noted that the retry parameters can be supplied to a service mesh (such as processing circuitry’ (cloud infrastructure 100) in Fig. 1A) by a configuration agent or utility. It is further noted that the retry parameters can be supplied as arguments to an invocation of a service function.

[0112] hi some embodiments, a configuration agent or utility utilizes different factors affecting retry priority (e.g, a.-c, above) to determine specific retry parameters that will be applied for particular invocations, and then provides those determined parameters to the microservices and / or service mesh (as appropriate).

[0113] Fig. 2 illustrates a flow diagram of an example method of differential retry after service invocation failure, in accordance with some embodiments of the presently disclosed subject matter.

[0114] The method illustrated in Fig. 2 can be utilized - for example - in a system based on the architecture shown in Fig. 1A. The method illustrated in Fig. 2 can be utilized ---for example - in a system based on the architecture shown in Fig. 1B-1C, with variations as noted herein.

[0115] Processing circuitry' (cloud infrastructure) 100 (e.g. differential retry' unit 115A) can obtain 205 a service invocation request originated by an invoking microservice.

[0116] In a system based on the architecture shown in Fig. 1A, processing circuitry (cloud infrastructure) 100 (e.g. differential retry unit 115) can obtain explicit or configured retry parameter data. In some embodiments, the retry parameter data accompanies the function invocation (for example: as shown above with reference to Fig.

[0117] 1A) — so that the retry parameter data is supplied by the invoking service. In some other embodiments, retry parameter data is configured e.g. at system initialization and / or realtime – as described above.

[0118] Retry parameter data can include, for example:

[0119] a maximum number of retries to be executed (e.g. in response to a given invocation)

[0120] a fixed between-retry delay

[0121] specification of a retry’ delay algorithm together with associated parameters By way of non-limiting example: the algorithm can be specified as exponential backoff, together with associated parameters such as initial backoff time and maximum backoff time. By way of further non-limiting example: the algorithm can be specified as a probability of retrying

[0122] specification of condition-based retry e.g. retry only if other higher -priority services have not recently been utilizing the invoked service

[0123] Retry parameter data can also include parameters that are dependent on microservice state, or system state. For example:

[0124] if a system load metric exceeds a certain value, then maximum number of retries is N0, otherwise the maximum number of retries is N1.

[0125] In some embodiments, retry parameters can be given as a priority ranking (e.g. bronze, silver, gold, or classes 1-5, or continuous ranking etc.). The priority ranking can then determine the particular retry behaviors (maximum number of retries etc.).

[0126] Processing circuitry (cloud infrastructure) 100 (e.g. differential retry unit 115A) can next attempt 210 to invoke the requested service.

[0127] Processing circuitry (cloud infrastructure) 100 (e.g. differential retry unit 115A) can next --- in a manner specific to the particular cloud provider - detect 215 success or failure of the invocation of the requested service.

[0128] Responsive to detection of failure of invocation, processing circuitry (cloud infrastructure) 100 (e.g. differential retry unit 115A) can utilize the retry' parameters to determine 220 whether to e.g. retry 225 - or schedule retrying - the sen-ice invocation.

[0129] In some embodiments, processing circuitry (cloud infrastructure) 100 (e.g.differential retry unit 115A) performs immediate conditional retry. For example, processing circuitry (cloud infrastructure) 100 (e.g. differential retry unit 115A) can, based on the retry parameters, examine:

[0130] system load (e.g. whether higher priority applications have recently utilized the invoked service)

[0131] number of retries attempted thus far

[0132] a randomly generated number (to be compared to a probability parameter) etc.

[0133] Processing circuitry (cloud infrastructure) 100 (e.g. differential retry unit 115A) can then retry (or refrain from retrying) based on this examined information.

[0134] In some embodiments, processing circuitry (cloud infrastructure) 100 (e.g. differential retry unit 115A) schedules retry. For example, processing circuitry' (cloud infrastructure) 100 (e.g. differential retry unit 115A) can, based on the retry' parameters, determine a time delay (e.g. 200 milliseconds) to wait until the retry. Processing circuitry(cloud infrastructure) 100 (e.g. differential retry unit 115A) can make the scheduled retry conditional - so that when the time is elapsed, processing circuitry (cloud infrastructure) 100 (e.g. differential retry unit 115A) can evaluate system load etc, to determine whether to perform the retry.

[0135] It is noted that processing circuitry (cloud infrastructure) 100 (e.g. differential retry unit 115A) can - based on retry parameters - combine conditional and scheduled retry i.e. processing circuitry (cloud infrastructure) 100 (e.g. differential retry unit 115A) can evaluate to whether to retry now, and if it determines not to retry now, it can schedule a future conditional or unconditional retry.

[0136] It is noted that in some embodiments, processing circuitry (cloud infrastructure) 100 (e.g. differential retry unit 115A) performs conditional invocation and / or scheduled invocation at the first invocation (and not only upon retry ).

[0137] It is noted that in some embodiments, processing circuitry (cloud infrastructure) 100 (e.g. differential retry unit 115A) performs conditional invocation and / or scheduled invocation regardless of current system load. It is noted that in some other embodiments, processing circuitry (cloud infrastructure) 100 (e.g. differential retry’ unit USA) performs conditional invocation and / or scheduled invocation in response to system load (e.g. a metric of system load meeting a threshold).

[0138] It is noted that in this manner resistance to denial-of-service attacks can be increased (and monetary cost resulting from such attacks can be reduced) - while maintaining service quality of applications (including high-priority applications).

[0139] It is further noted that the method of Fig. 2 can be utilized in systems such as those depicted in Fig. IB and Fig. 1C. In these systems, the m ethod of Fig. 2 can be performed by e.g. Service_1 120 (e.g. differential retry unit 115B 115C)

[0140] Fig. 3 illustrates a logical block diagram of an example application, together with its utilized cloud-based microservice, with enablement of enhanced load tolerance (e.g. for denial-of-service resistance), in accordance with some embodiments of the presently disclosed subject matter.

[0141] Service 1 305 can be a microservice executing in cloud infrastructure 310, Service 1 305 can include denial-of-service (DoS) risk scoring unit 320. Service 1 305 can include application data (cookie) analysis unit 335. Service_l 305 can receive e.g. hypertext transfer protocol (HTTP) data via Internet 315 from, for example, HTTP-based application 300.

[0142] By way of non-limiting example, Service_l 305 can be an application which exchanges data with clients using HTTP (e.g. HTTP-based Representational State Transfer (REST) messages). The HTTP data can include e.g. cookies or other data that identifies users as well as user history.

[0143] Service_l 305 (for example, application data (cookie) analysis unit 335) can utilize HTTP data and / or application state to prioritize users and / or applications.

[0144] By way of non-limiting example: Service_l 305 can determine a priority ranking based on application-specific characteristics (e.g. presence or amount of merchandise in a shopping basket, session duration or idle time, user history, frequency or user visits, frequency of purchases, service visited so far, etc. ) as determined from cookies and / or other HTTP data. The determined ranking can be indicative of retry parameters (maximum retries etc. as described herein above). Alternatively, Service_l 305 can determine retry parameters directly based on the application-specific characteristics.

[0145] Service 1 305 (for example, DoS risk scoring unit 320) can utilize HTTP data or application state to determine a score indicating retry priority and / or retry parameters for avoidance of denial-of-service.

[0146] By way of non-limiting example: Service 1 305 can determine a risk score based on an estimation of likelihood of a particular client being originator of a denial-of-service attack aimed at the cloud infrastructure. The determined risk score can be indicative of retry' parameters (maximum retries etc. as described herein above). Alternatively,

[0147] Service 1 305 can determine retry parameters directly based on the estimation of likelihood of a particular client being originator of a denial-of-service attack aimed at e.g. the cloud infrastructure.

[0148] Fig. 4 illustrates a flow diagram of an example method of prioritization of cloudbased service utilization based on information derived from client data (e.g. cookies), in accordance with some embodiments of the presently disclosed subject matter.

[0149] Processing circuitry (cloud infrastructure) 310 (e.g. Service 1 305) can receive 405 data e.g. hypertext transfer protocol (HTTP) data via Internet 315 from a client (for example: HTTP-based application 300).

[0150] Processing circuitry’ (cloud infrastructure) 310 (for example: DoS risk scoring unit 320) can then derive 410 - from the received data (or application state information derivative of the received data) - data indicative of client / application priority and / or service invocation retry parameters. For example: processing circuitry (cloud infrastructure) 310 (for example: DoS risk scoring unit 320) can derive a denial-of-service (DoS) risk score, or a client priority ranking.

[0151] Processing circuitry (cloud infrastructure) 310 (for example: DoS risk scoring unit 320) can - for example - derive a high DoS risk score after identifying a data source (e.g. based on cookie data) as being a previously unseen user. Service 1 305 (for example: DoS risk scoring unit 320) can similarly derive a low’ DoS risk score after identifying a data source (e.g. based on cookie data) as — for example, a paying user. Service_l 305 can utilize this HTTP-derived risk score in application security. By way of non-limiting example, Service_l 305 (for example: DoS risk scoring unit 320) can give high priority to retrying invocations of microservices, where the invocations are associated w ith clients that have a low’ denial-of-service risk score. Conversely, Service 1 305 (for example: DoS risk: scoring unit 320) can give low priority to retiying invocations of microservices, where the invocations are associated with clients that have a high denial-of-service risk score.

[0152] Processing circuitry (cloud infrastructure) 310 (for example: DoS risk scoring unit 320) can - for example - derive a high client priority’ ranking after identifying a data source (e.g. based on cookie data) as being a high frequency user. Service l 305 (for example: DoS risk scoring unit 320) can similarly derive a low client priority ranking after identifying a data source (e.g. based on cookie data) as a user who visits frequently but never completes purchases, Service_l 305 (for example: DoS risk scoring unit 320)can then give high priority to retrying invocations of microservices, where the invocations are associated with clients that have a high client priority ranking.

[0153] Conversely, Service_l 305 (for example: DoS risk scoring unit 320) can give low priority to retrying invocations of microservices, where the invocations are associated with clients that have a low client priority ranking.

[0154] Processing circuitry (cloud infrastructure) 310 (for example: Service_l 305) can next, based on e.g. the received data, communicate with processing circuitry (cloud infrastructure) 310 to invoke 415 a function provided by a second microservice (second microservice is not shown). Processing circuitry (cloud infrastructure) 310 (for example: Service_1 305) can (in some embodiments) supply data indicative of retry parameters as part of the invocation. Processing circuitry (cloud infrastructure) 310 can (in some other embodiments) supply data indicative of retry parameters through a different mechanism, as described hereinabove. Processing circuitry (cloud infrastructure) 310 can then utilize 420 these parameters as part of differential retry, as described above with reference to Fig. 2.

[0155] In some embodiments, processing circuitry (cloud infrastructure) 310 (for example: Service_1 305) can further, based on e.g. the received data, prioritize a first request with a higher user or application (service) priority over a second request with a lesser higher user or application (service) priority. For example: processing circuitry (cloud infrastructure) 310 (for example: Service_l 305) can invoke a microservice associated with a higher user or application (service) priority before a microservice invocation associated with a lower user or application (service) priority.

[0156] Those skilled in the art will readily appreciate that various modifications and changes can be applied to the embodiments of the invention as hereinbefore described without departing from its scope, defined in and by the appended claims.

Claims

CLAIMS1. A system of cloud-based interservice invocation, the system comprising a processing circuitry (PC) configured to:a) invoke the service function provided by a first cloud service; andb) responsive to a failure of the invoking of the service function, perform at least one of:i) scheduling a reinvocation of the service function based on the one or more service invocation retry parameters associated with the invoking, orii) conditionally reinvoking the service function, based on the one or more service invocation retry parameters associated with the invoking.

2. The system of claim 1, wherein a) - b) are performed by a service mesh, and wherein the PC is additionally configured to perform, prior to a):obtain a request of invocation, by a second cloud service, of the service function provided by the first cloud service.

3. The system of claim 1, wherein at least one of the one or more service invocation retry parameters is derivative of at least one of:a. an identity of an invoking microservice;b. data determined by an invoking microservice or predecessor invoking microservice; andc. data derivative of the identity’ of the client system.

4. The system of claim 3, wherein the data determined by an invoking microservice or predecessor invoking microservice comprises a client priority ranking or deniai-of-service risk score.

5. The system of claim 3, wherein the data derivative of the identity’ of the client sy stem is based on network monitoring.

6. The system of claim 1, wherein the service invocation retry parameters comprise at least one of:i) a between-retry delay value,ii) a maximum retry count,iii) a probability of retrying, oriv) a maximum value of a randomized between-retry delay value.

7. A processing-circuitry-based method of cloud-based interservice invocation, the system comprising a processing circuitry (PC) configured to:a) invoking the service function provided by a first cloud service: andb) responsive to a failure of the invoking of the service function, performing at least one of:i) scheduling a reinvocation of the service function based on the one or more service invocation retry parameters associated with the invoking, orii) conditionally reinvoking the service function, based on the one or more service invocation retry parameters associated with the invoking.

8. A computer program product comprising a computer readable non-transitory storage medium containing program instructions, which program instructions when read by a processor, cause the processing circuitry to perform a method of cloud-based interservice invocation, the method comprising:a) invoking the service function provided by a first cloud service: andb) responsive to a failure of the invoking of the service function, performing at least one of:i) scheduling a reinvocation of the service function based on the one or more service invocation retry parameters associated with the invoking, orii) conditionally reinvoking the service function, based on the one or more service invocation retry parameters associated with the invoking.

9. A system of cloud-based services enabling enhanced load tolerance, the system comprising a processing circuitry (PC) configured to:receive, by a first cloud service, data originated by a client;determine, based on the data originated by the client, data indicative of service invocation retry parameters;invoke a service function provided by a second microservice; andresponsive to a failure of the invoking of the service function, perform retry of the invoking, based on the service invocation retry parameters.

10. The system of claim 9, wherein the PC is further configured to:identify cookie data in the data originated by the client; andutilize, at least, the identified cookie data in the determination of the service invocation retry parameters.

11. The system of claim 9, wherein the PC is further configured to:determine a denial-of service risk score based on data originated by the client; and utilize, at least, the denial-of-service risk score in the determination of the service invocation retry parameters.

12. The system of claim 9, wherein the PC is further configured to:determine a client priority ranking based on data originated by the client; and utilize, at least, the client priority ranking in the determination of the service invocation retry parameters.

13. A processing-circuitry-based method of enabling enhanced load tolerance in cloud-based services, the method comprising:receiving, by a first cloud service, data originated by a client; determining, based on the data originated by the client, data indicative of service invocation retry parameters;invoking a service function provided by a second microservice; and responsive to a failure of the invoking of the service function, performing retry of the invoking, based on the service invocation retry parameters.

14. A computer program product comprising a computer readable non-transitory storage medium containing program instructions, which program instructions when read bya processor, cause the processing circuitry’ to perform a method of enabling enhanced load tolerance in cloud-based sendees, the method comprising:receiving, by a first cloud service, data originated by a client; determining, based on the data originated by the client, data indicative of service invocation retry parameters;invoking a service function provided by a second microservice; and responsive to a failure of the invoking of the service function, performing retry of the invoking, based on the service invocation retry parameters,