Inception score

The Inception Score (IS) is an algorithm used to assess the quality of images created by a generative image model such as a generative adversarial network (GAN). The score is calculated based on the output of a separate, pretrained Inception v3 image classification model applied to a sample of (typically around 30,000) images generated by the generative model. The Inception Score is maximized when the following conditions are true: The entropy of the distribution of labels predicted by the Inceptionv3 model for the generated images is minimized. In other words, the classification model confidently predicts a single label for each image. Intuitively, this corresponds to the desideratum of generated images being "sharp" or "distinct". The predictions of the classification model are evenly distributed across all possible labels. This corresponds to the desideratum that the output of the generative model is "diverse". It has been somewhat superseded by the related Fréchet inception distance. While the Inception Score only evaluates the distribution of generated images, the FID compares the distribution of generated images with the distribution of a set of real images ("ground truth"). == Definition == Let there be two spaces, the space of images Ω X {\displaystyle \Omega _{X}} and the space of labels Ω Y {\displaystyle \Omega _{Y}} . The space of labels is finite. Let p g e n {\displaystyle p_{gen}} be a probability distribution over Ω X {\displaystyle \Omega _{X}} that we wish to judge. Let a discriminator be a function of type p d i s : Ω X → M ( Ω Y ) {\displaystyle p_{dis}:\Omega _{X}\to M(\Omega _{Y})} where M ( Ω Y ) {\displaystyle M(\Omega _{Y})} is the set of all probability distributions on Ω Y {\displaystyle \Omega _{Y}} . For any image x {\displaystyle x} , and any label y {\displaystyle y} , let p d i s ( y | x ) {\displaystyle p_{dis}(y|x)} be the probability that image x {\displaystyle x} has label y {\displaystyle y} , according to the discriminator. It is usually implemented as an Inception-v3 network trained on ImageNet. The Inception Score of p g e n {\displaystyle p_{gen}} relative to p d i s {\displaystyle p_{dis}} is I S ( p g e n , p d i s ) := exp ⁡ ( E x ∼ p g e n [ D K L ( p d i s ( ⋅ | x ) ‖ ∫ p d i s ( ⋅ | x ) p g e n ( x ) d x ) ] ) {\displaystyle IS(p_{gen},p_{dis}):=\exp \left(\mathbb {E} _{x\sim p_{gen}}\left[D_{KL}\left(p_{dis}(\cdot |x)\|\int p_{dis}(\cdot |x)p_{gen}(x)dx\right)\right]\right)} Equivalent rewrites include ln ⁡ I S ( p g e n , p d i s ) := E x ∼ p g e n [ D K L ( p d i s ( ⋅ | x ) ‖ E x ∼ p g e n [ p d i s ( ⋅ | x ) ] ) ] {\displaystyle \ln IS(p_{gen},p_{dis}):=\mathbb {E} _{x\sim p_{gen}}\left[D_{KL}\left(p_{dis}(\cdot |x)\|\mathbb {E} _{x\sim p_{gen}}[p_{dis}(\cdot |x)]\right)\right]} ln ⁡ I S ( p g e n , p d i s ) := H [ E x ∼ p g e n [ p d i s ( ⋅ | x ) ] ] − E x ∼ p g e n [ H [ p d i s ( ⋅ | x ) ] ] {\displaystyle \ln IS(p_{gen},p_{dis}):=H[\mathbb {E} _{x\sim p_{gen}}[p_{dis}(\cdot |x)]]-\mathbb {E} _{x\sim p_{gen}}[H[p_{dis}(\cdot |x)]]} ln ⁡ I S {\displaystyle \ln IS} is nonnegative by Jensen's inequality. Pseudocode:INPUT discriminator p d i s {\displaystyle p_{dis}} . INPUT generator g {\displaystyle g} . Sample images x i {\displaystyle x_{i}} from generator. Compute p d i s ( ⋅ | x i ) {\displaystyle p_{dis}(\cdot |x_{i})} , the probability distribution over labels conditional on image x i {\displaystyle x_{i}} . Sum up the results to obtain p ^ {\displaystyle {\hat {p}}} , an empirical estimate of ∫ p d i s ( ⋅ | x ) p g e n ( x ) d x {\displaystyle \int p_{dis}(\cdot |x)p_{gen}(x)dx} . Sample more images x i {\displaystyle x_{i}} from generator, and for each, compute D K L ( p d i s ( ⋅ | x i ) ‖ p ^ ) {\displaystyle D_{KL}\left(p_{dis}(\cdot |x_{i})\|{\hat {p}}\right)} . Average the results, and take its exponential. RETURN the result. === Interpretation === A higher inception score is interpreted as "better", as it means that p g e n {\displaystyle p_{gen}} is a "sharp and distinct" collection of pictures. ln ⁡ I S ( p g e n , p d i s ) ∈ [ 0 , ln ⁡ N ] {\displaystyle \ln IS(p_{gen},p_{dis})\in [0,\ln N]} , where N {\displaystyle N} is the total number of possible labels. ln ⁡ I S ( p g e n , p d i s ) = 0 {\displaystyle \ln IS(p_{gen},p_{dis})=0} iff for almost all x ∼ p g e n {\displaystyle x\sim p_{gen}} p d i s ( ⋅ | x ) = ∫ p d i s ( ⋅ | x ) p g e n ( x ) d x {\displaystyle p_{dis}(\cdot |x)=\int p_{dis}(\cdot |x)p_{gen}(x)dx} That means p g e n {\displaystyle p_{gen}} is completely "indistinct". That is, for any image x {\displaystyle x} sampled from p g e n {\displaystyle p_{gen}} , discriminator returns exactly the same label predictions p d i s ( ⋅ | x ) {\displaystyle p_{dis}(\cdot |x)} . The highest inception score N {\displaystyle N} is achieved if and only if the two conditions are both true: For almost all x ∼ p g e n {\displaystyle x\sim p_{gen}} , the distribution p d i s ( y | x ) {\displaystyle p_{dis}(y|x)} is concentrated on one label. That is, H y [ p d i s ( y | x ) ] = 0 {\displaystyle H_{y}[p_{dis}(y|x)]=0} . That is, every image sampled from p g e n {\displaystyle p_{gen}} is exactly classified by the discriminator. For every label y {\displaystyle y} , the proportion of generated images labelled as y {\displaystyle y} is exactly E x ∼ p g e n [ p d i s ( y | x ) ] = 1 N {\displaystyle \mathbb {E} _{x\sim p_{gen}}[p_{dis}(y|x)]={\frac {1}{N}}} . That is, the generated images are equally distributed over all labels.

Human visual system model

A human visual system model (HVS model) is used by image processing, video processing and computer vision experts to deal with biological and psychological processes that are not yet fully understood. Such a model is used to simplify the behaviors of what is a very complex system. As our knowledge of the true visual system improves, the model is updated. Psychovisual study is the study of the psychology of vision. The human visual system model can produce desired effects in perception and vision. Examples of using an HVS model include color television, lossy compression, and Cathode-ray tube (CRT) television. Originally, it was thought that color television required too high a bandwidth for the then available technology. Then it was noticed that the color resolution of the HVS was much lower than the brightness resolution; this allowed color to be squeezed into the signal by chroma subsampling. Another example is lossy image compression, like JPEG. Our HVS model says we cannot see high frequency detail, so in JPEG we can quantize these components without a perceptible loss of quality. Similar concepts are applied in audio compression, where sound frequencies inaudible to humans are band-stop filtered. Several HVS features are derived from evolution when we needed to defend ourselves or hunt for food. We often see demonstrations of HVS features when we are looking at optical illusions. == Block diagram of HVS == == Assumptions about the HVS == Low-pass filter characteristic (limited number of rods in human eye): see Mach bands Lack of color resolution (fewer cones in human eye than rods) Motion sensitivity More sensitive in peripheral vision Stronger than texture sensitivity, e.g. viewing a camouflaged animal Texture stronger than disparity – 3D depth resolution does not need to be so accurate Integral Face recognition (babies smile at faces) Depth inverted face looks normal (facial features overrule depth information) Upside down face with inverted mouth and eyes looks normal == Examples of taking advantage of an HVS model == Flicker frequency of film and television using persistence of vision to fool viewer into seeing a continuous image Interlaced television painting half images to give the impression of a higher flicker frequency Color television (chrominance at half resolution of luminance corresponding to proportions of rods and cones in eye) Image compression (difficult to see higher frequencies more harshly quantized) Motion estimation (use luminance and ignore color) Watermarking and Steganography

List of cryptography journals

List of cryptography journals includes notable peer-reviewed academic journals that focus on cryptography, cryptanalysis, information security, and related areas in computer science and mathematics. == Notable journals == Cryptologia Designs, Codes and Cryptography IEEE Transactions on Information Theory International Journal of Information Security Journal of Cryptology Journal of Computer Security ACM Transactions on Privacy and Security Information Processing Letters Information and Computation

Trusted Computing

Trusted Computing (TC) is a technology developed and promoted by the Trusted Computing Group. The term is taken from the field of trusted systems and has a specialized meaning that is distinct from the field of confidential computing. With Trusted Computing, the computer will consistently behave in expected ways, and those behaviors will be enforced by computer hardware and software. Enforcing this behavior is achieved by loading the hardware with a unique encryption key that is inaccessible to the rest of the system and the owner. TC is controversial as the hardware is not only secured for its owner, but also against its owner, leading opponents of the technology like free software activist Richard Stallman to deride it as "treacherous computing", and certain scholarly articles to use scare quotes when referring to the technology. Trusted Computing proponents such as International Data Corporation, the Enterprise Strategy Group and Endpoint Technologies Associates state that the technology will make computers safer, less prone to viruses and malware, and thus more reliable from an end-user perspective. They also state that Trusted Computing will allow computers and servers to offer improved computer security over that which is currently available. Opponents often state that this technology will be used primarily to enforce digital rights management policies (imposed restrictions to the owner) and not to increase computer security. Chip manufacturers Intel and AMD, hardware manufacturers such as HP and Dell, and operating system providers such as Microsoft include Trusted Computing in their products if enabled. The U.S. Army requires that every new PC it purchases comes with a Trusted Platform Module (TPM). As of July 3, 2007, so does virtually the entire United States Department of Defense. == Key concepts == Trusted Computing encompasses six key technology concepts, of which all are required for a fully Trusted system, that is, a system compliant to the TCG specifications: Endorsement key Secure input and output Memory curtaining / protected execution Sealed storage Remote attestation Trusted Third Party (TTP) === Endorsement key === The endorsement key is a 2048-bit RSA public and private key pair that is created randomly on the chip at manufacture time and cannot be changed. The private key never leaves the chip, while the public key is used for attestation and for encryption of sensitive data sent to the chip, as occurs during the TPM_TakeOwnership command. This key is used to allow the execution of secure transactions: every Trusted Platform Module (TPM) is required to be able to sign a random number (in order to allow the owner to show that he has a genuine trusted computer), using a particular protocol created by the Trusted Computing Group (the direct anonymous attestation protocol) in order to ensure its compliance of the TCG standard and to prove its identity; this makes it impossible for a software TPM emulator with an untrusted endorsement key (for example, a self-generated one) to start a secure transaction with a trusted entity. The TPM should be designed to make the extraction of this key by hardware analysis hard, but tamper resistance is not a strong requirement. === Memory curtaining === Memory curtaining extends common memory protection techniques to provide full isolation of sensitive areas of memory—for example, locations containing cryptographic keys. Even the operating system does not have full access to curtained memory. The exact implementation details are vendor specific. === Sealed storage === Sealed storage protects private information by binding it to platform configuration information including the software and hardware being used. This means the data can be released only to a particular combination of software and hardware. Sealed storage can be used for DRM enforcing. For example, users who keep a song on their computer that has not been licensed to be listened will not be able to play it. Currently, a user can locate the song, listen to it, and send it to someone else, play it in the software of their choice, or back it up (and in some cases, use circumvention software to decrypt it). Alternatively, the user may use software to modify the operating system's DRM routines to have it leak the song data once, say, a temporary license was acquired. Using sealed storage, the song is securely encrypted using a key bound to the trusted platform module so that only the unmodified and untampered music player on his or her computer can play it. In this DRM architecture, this might also prevent people from listening to the song after buying a new computer, or upgrading parts of their current one, except after explicit permission of the vendor of the song. === Remote attestation === Remote attestation allows changes to the user's computer to be detected by authorized parties. For example, software companies can identify unauthorized changes to software, including users modifying their software to circumvent commercial digital rights restrictions. It works by having the hardware generate a certificate stating what software is currently running. The computer can then present this certificate to a remote party to show that unaltered software is currently executing. Numerous remote attestation schemes have been proposed for various computer architectures, including Intel, RISC-V, and ARM. Remote attestation is usually combined with public-key encryption so that the information sent can only be read by the programs that requested the attestation, and not by an eavesdropper. To take the song example again, the user's music player software could send the song to other machines, but only if they could attest that they were running an authorized copy of the music player software. Combined with the other technologies, this provides a more restricted path for the music: encrypted I/O prevents the user from recording it as it is transmitted to the audio subsystem, memory locking prevents it from being dumped to regular disk files as it is being worked on, sealed storage curtails unauthorized access to it when saved to the hard drive, and remote attestation prevents unauthorized software from accessing the song even when it is used on other computers. To preserve the privacy of attestation responders, Direct Anonymous Attestation has been proposed as a solution, which uses a group signature scheme to prevent revealing the identity of individual signers. Proof of space (PoS) have been proposed to be used for malware detection, by determining whether the L1 cache of a processor is empty (e.g., has enough space to evaluate the PoSpace routine without cache misses) or contains a routine that resisted being evicted. === Trusted third party === == Known applications == The Microsoft products Windows Vista, Windows 7, Windows 8 and Windows RT make use of a Trusted Platform Module to facilitate BitLocker Drive Encryption. Other known applications with runtime encryption and the use of secure enclaves include the Signal messenger and the e-prescription service ("E-Rezept") by the German government. == Possible applications == === Digital rights management === Trusted Computing would allow companies to create a digital rights management (DRM) system which would be very hard to circumvent, though not impossible. An example is downloading a music file. Sealed storage could be used to prevent the user from opening the file with an unauthorized player or computer. Remote attestation could be used to authorize play only by music players that enforce the record company's rules. The music would be played from curtained memory, which would prevent the user from making an unrestricted copy of the file while it is playing, and secure I/O would prevent capturing what is being sent to the sound system. Circumventing such a system would require either manipulation of the computer's hardware, capturing the analogue (and thus degraded) signal using a recording device or a microphone, or breaking the security of the system. New business models for use of software (services) over Internet may be boosted by the technology. By strengthening the DRM system, one could base a business model on renting programs for a specific time periods or "pay as you go" models. For instance, one could download a music file which could only be played a certain number of times before it becomes unusable, or the music file could be used only within a certain time period. === Preventing cheating in online games === Trusted Computing could be used to combat cheating in online games. Some players modify their game copy in order to gain unfair advantages in the game; remote attestation, secure I/O and memory curtaining could be used to determine that all players connected to a server were running an unmodified copy of the software. === Verification of remote computation for grid computing === Trusted Computing could be used to guarantee participants in a grid computing sys

Social network hosting service

A social network hosting service is a web hosting service that specifically hosts the user creation of web-based social networking services, alongside related applications. Such services are also known as vertical social networks due to the creation of SNSes which cater to specific user interests and niches; like larger, interest-agnostic SNSes, such niche networking services may also possess the ability to create increasingly niche groups of users. == List of social network hosting services == Federated Media Publishing's BigTent BroadVision Clearvale Ning Wall.fm

Algorithmic bias

Algorithmic bias describes systematic and repeatable harmful tendency in a computerized sociotechnical system to create "unfair" outcomes, such as "privileging" one category over another in ways that may or may not be different from the intended function of the algorithm. Bias can emerge from many factors, including intentionally biased design decisions or the unintended or unanticipated use or decisions relating to the way data is coded, collected, selected or used to train the algorithm. For example, algorithmic bias has been observed in search engine results and social media platforms. This bias can have impacts ranging from privacy violations to reinforcing social biases of race, gender, sexuality, and ethnicity. The study of algorithmic bias is most concerned with algorithms that reflect "systematic and unfair" discrimination. This bias has only recently been addressed in legal frameworks, such as the European Union's General Data Protection Regulation (enforced in 2018) and the Artificial Intelligence Act (proposed in 2021 and adopted in 2024). As algorithms expand their ability to organize society, politics, institutions, and behavior, sociologists have become concerned with the ways in which unanticipated output and manipulation of data can impact the physical world. Because algorithms are often considered to be neutral and unbiased, they can inaccurately project greater authority than human expertise (in part due to the psychological phenomenon of automation bias), and in some cases, reliance on algorithms can displace human responsibility for their outcomes, without last mile thinking. Bias can enter into algorithmic systems as a result of pre-existing cultural, social, or institutional expectations; by how features and labels are chosen; because of technical limitations of their design; or by being used in unanticipated contexts or by audiences who are not considered in the software's initial design. Algorithmic bias has been cited in cases ranging from election outcomes to the spread of online hate speech. It has also arisen in criminal justice, healthcare, and hiring, compounding existing racial, socioeconomic, and gender biases. The relative inability of facial recognition technology to accurately identify darker-skinned faces has been linked to multiple wrongful arrests of black men, an issue stemming from imbalanced datasets. Problems in understanding, researching, and discovering algorithmic bias persist due to the proprietary nature of algorithms, which are typically treated as trade secrets. Even when full transparency is provided, the complexity of certain algorithms poses a barrier to understanding their functioning. Furthermore, algorithms may change, or respond to input or output in ways that cannot be anticipated or easily reproduced for analysis. In many cases, even within a single website or application, there is no single "algorithm" to examine, but a network of many interrelated programs and data inputs, even between users of the same service. A 2021 survey identified multiple forms of algorithmic bias, including historical, representation, and measurement biases, each of which can contribute to unfair outcomes. == Definitions == Algorithms are difficult to define, but may be generally understood as lists of instructions that determine how programs read, collect, process, and analyze data to generate a usable output. For a rigorous technical introduction, see Algorithms. Advances in computer hardware and software have led to an increased capability to process, store and transmit data. This has in turn made the design and adoption of technologies such as machine learning and artificial intelligence technically and commercially feasible. By analyzing and processing data, algorithms are the backbone of search engines, social media websites, recommendation engines, online retail, online advertising, and more. Contemporary social scientists are concerned with algorithmic processes embedded into hardware and software applications because of their political and social impact, and question the underlying assumptions of an algorithm's neutrality. The term algorithmic bias describes systematic and repeatable errors that create unfair outcomes, such as privileging one arbitrary group of users over others. For example, a credit score algorithm may deny a loan without being unfair, if it is consistently weighing relevant financial criteria. If the algorithm recommends loans to one group of users, but denies loans to another set of nearly identical users based on unrelated criteria, and if this behavior can be repeated across multiple occurrences, an algorithm can be described as biased. This bias may be intentional or unintentional (for example, it can come from biased data obtained from a worker that previously did the job the algorithm is going to do from now on). == Methods == Bias can be introduced to an algorithm in several ways. During the assemblage of a dataset, data may be collected, digitized, adapted, and entered into a database according to human-designed cataloging criteria. Next, programmers assign priorities, or hierarchies, for how a program assesses and sorts that data. This requires human decisions about how data is categorized, and which data is included or discarded. Some algorithms collect their own data based on human-selected criteria, which can also reflect the bias of human designers. Other algorithms may reinforce stereotypes and preferences as they process and display "relevant" data for human users, for example, by selecting information based on previous choices of a similar user or group of users. Beyond assembling and processing data, bias can emerge as a result of design. For example, algorithms that determine the allocation of resources or scrutiny (such as determining school placements) may inadvertently discriminate against a category when determining risk based on similar users (as in credit scores). Meanwhile, recommendation engines that work by associating users with similar users, or that make use of inferred marketing traits, might rely on inaccurate associations that reflect broad ethnic, gender, socio-economic, or racial stereotypes. Another example comes from determining criteria for what is included and excluded from results. These criteria could present unanticipated outcomes for search results, such as with flight-recommendation software that omits flights that do not follow the sponsoring airline's flight paths. Algorithms may also display an uncertainty bias, offering more confident assessments when larger data sets are available. This can skew algorithmic processes toward results that more closely correspond with larger samples, which may disregard data from underrepresented populations. == History == === Early critiques === The earliest computer programs were designed to mimic human reasoning and deductions, and were deemed to be functioning when they successfully and consistently reproduced that human logic. In his 1976 book Computer Power and Human Reason, artificial intelligence pioneer Joseph Weizenbaum suggested that bias could arise both from the data used in a program, but also from the way a program is coded. Weizenbaum wrote that programs are a sequence of rules created by humans for a computer to follow. By following those rules consistently, such programs "embody law", that is, enforce a specific way to solve problems. The rules a computer follows are based on the assumptions of a computer programmer for how these problems might be solved. That means the code could incorporate the programmer's imagination of how the world works, including their biases and expectations. While a computer program can incorporate bias in this way, Weizenbaum also noted that any data fed to a machine additionally reflects "human decision making processes" as data is being selected. Finally, he noted that machines might also transfer good information with unintended consequences if users are unclear about how to interpret the results. Weizenbaum warned against trusting decisions made by computer programs that a user doesn't understand, comparing such faith to a tourist who can find his way to a hotel room exclusively by turning left or right on a coin toss. Crucially, the tourist has no basis of understanding how or why he arrived at his destination, and a successful arrival does not mean the process is accurate or reliable. An early example of algorithmic bias resulted in as many as 60 women and ethnic minorities denied entry to St. George's Hospital Medical School per year from 1982 to 1986, based on implementation of a new computer-guidance assessment system that denied entry to women and men with "foreign-sounding names" based on historical trends in admissions. While many schools at the time employed similar biases in their selection process, St. George was most notable for automating said bias through the use of an algorithm, thus gaining the attention of people on a much

Cloud Data Management Interface

ISO/IEC 17826 Information technology — Cloud Data Management Interface (CDMI) Version 2.0.0 is an international standard that specifies a protocol for self-provisioning, administering and managing access to data stored in cloud storage, object storage, storage area network and network attached storage systems. The CDMI standard is developed and maintained by the Storage Networking Industry Association, who makes a publicly accessible version of the specification available. CDMI defines new resource representations to enable standardized management of any URI-accessible data, and defines RESTful HTTP operations using these representations to discover the capabilities of the storage system, discover stored data, access and update management metadata, specify data storage protocols (such as iSCSI and NFS) through which the stored data is accessed, and provide cross-system and cross-cloud import and export in order to enable data portability. Management functions enabled by CDMI include managing data ownership, identity mapping, access controls, user-specified metadata, and to declaratively specify desired data protection, data retention, constraints on geographic placement, desired quality of service, data versioning and security requirements. CDMI also defines utility services to facilitate data management, such the ability to query data matching specific criteria, and includes extensions to perform bulk updates using CDMI Jobs. == Capabilities == Compliant implementations must provide access to a set of configuration parameters known as capabilities. These are either boolean values that represent whether or not a system supports things such as queues, export via other protocols, path-based storage and so on, or numeric values expressing system limits, such as how much metadata may be placed on an object. As a minimal compliant implementation can be quite small, with few features, clients need to check the cloud storage system for a capability before attempting to use the functionality it represents. Resource allocation assignments limited to the data management interface protocols must possess access bypass capabilities which extend beyond the layered framework. This integral function is vital to the prevention of transport layer session hijacking by unauthorized entities which may circumvent standard interfacing security parameters. == Containers == A CDMI client may access objects, including containers, by either name or object id (OID), assuming the CDMI server supports both methods. When storing objects by name, it is natural to use nested named containers; the resulting structure corresponds exactly to a traditional filesystem directory structure. == Objects == Objects are similar to files in a traditional file system, but are enhanced with an increased amount and capacity for metadata. As with containers, they may be accessed by either name or OID. When accessed by name, clients use URLs that contain the full pathname of objects to create, read, update and delete them. When accessed by OID, the URL specifies an OID string in the cdmi-objectid container; this container presents a flat name space conformant with standard object storage system semantics. Subject to system limits, objects may be of any size or type and have arbitrary user-supplied metadata attached to them. Systems that support query allow arbitrary queries to be run against the metadata. == Domains, Users and Groups == CDMI supports the concept of a domain, similar in concept to a domain in the Windows Active Directory model. Users and groups created in a domain share a common administrative database and are known to each other on a "first name" basis, i.e. without reference to any other domain or system. Domains also function as containers for usage and billing summary data. == Access Control == CDMI exactly follows the ACL and ACE model used for file authorization operations by NFSv4. This makes it also compatible with Microsoft Windows systems. == Metadata == CDMI draws much of its metadata model from the XAM specification. Objects and containers have "storage system metadata", "data system metadata" and arbitrary user specified metadata, in addition to the metadata maintained by an ordinary filesystem (atime etc.). == Queries == CDMI specifies a way for systems to support arbitrary queries against CDMI containers, with a rich set of comparison operators, including support for regular expressions. == Queues == CDMI supports the concept of persistent FIFO (first-in, first-out) queues. These are useful for job scheduling, order processing and other tasks in which lists of things must be processed in order. == Compliance == Both retention intervals and retention holds are supported by CDMI. A retention interval consists of a start time and a retention period. During this time interval, objects are preserved as immutable and may not be deleted. A retention hold is usually placed on an object because of judicial action and has the same effect: objects may not be changed nor deleted until all holds placed on them are removed. == Billing == Summary information suitable for billing clients for on-demand services can be obtained by authorized users from systems that support it. == Serialization == Serialization of objects and containers allows export of all data and metadata on a system and importation of that data into another cloud system. == Foreign protocols == CDMI supports export of containers as NFS or CIFS shares. Clients that mount these shares see the container hierarchy as an ordinary filesystem directory hierarchy, and the objects in the containers as normal files. Metadata outside of ordinary filesystem metadata may or may not be exposed. Provisioning of iSCSI LUNs is also supported. == Client SDKs == CDMI Reference Implementation Droplet libcdmi-java libcdmi-python .NET SDK