The Stack On Software and Sovereignty

1. Stack as Model

Stacks are a kind of platform that also happens to be structured through vertical interoperable layers, both hard and soft, global and local. Its properties are generic, extensible, and pliable; it provides modular recombinancy but only within the bounded set of its synthetic planes. It is an autogenerative parametric topography, but one that grows precisely through an initial subdivision of technologies into planar layers and then through an autocratic consolidation and rationalization of these through internal interfaces and protocols. As for any platform, that consolidation is driven less from centrally planned legal prescription than through the algorithmic conduction of selfdirected behaviors by free-range Users. The Stack discussed in the following chapters is a vast software/hardware formation, a proto-megastructure built of crisscrossed oceans, layered concrete and fiber optics, urban metal and fleshy fingers, abstract identities and the fortified skins of oversubscribed national sovereignty. It is a machine literally circumscribing the planet, which not only pierces and distorts Westphalian models of state territory but also produces new spaces in its own image: clouds, networks, zones, social graphs, ecologies, megacities, formal and informal violence, weird theologies, all superimposed one on the other. This aggregate machine becomes a systematic technology according to the properties and limitations of that very spatial order. The layers of The Stack, some continental in scale and others microscopic, work in specific relation to the layer above and below it. As I have suggested, the fragile complementarity between the layers composing The Stack is discussed both as an idealized model for how platforms may be designed and as a description of some of the ways that they already work now. The metaphor and the machine are diagrams made real in the megastructure.

If you start looking for them, “ stacks ” are everywhere. In a way, the Earth itself is a spherical stack, from its molten core, to the lower and upper mantle, to the crust on which organic life evolved under the troposphere, stratosphere, mesosphere, thermosphere, and exosphere. Humans evolved between two and only two of these layers. Charles and Ray Eames ’ s famous “ Powers of Ten ” films for IBM showed generations of high school students how to start from one everyday spot and from there think down to $1 0 ^ { - 9 }$ meters and up to $1 0 ^ { 2 3 }$ meters, from quarks to walls of galaxies, and back again. In a way, their presentation is a kind of telescoping stack. Archaeology organizes and depicts the temporality of unearthed assemblages according to the Harris matrix, and its interlocking principles of original horizontality, original continuity, and stratigraphic succession. The Marxian model of base and superstructure provided another verticalized image of social totality, whereby economic structural causality flows bottom-up, from foundational technical processes of production, valuation, and relations in the base, to their ultimate expression in cultural and political institutions, as superstructure. Marx wanted to model historical cause and effect, but history is full of images of society organized instead into static stratified layers of arbitrary hierarchies (Albrecht D ü rer ’ s 1515 woodcut The Triumphal Arch of Maximilian I comes to mind). Many contemporary technical systems work on stack principles, including smart grids that segment a power layer, below a communications layer, below optimization and applications layers. Examples are plentiful, and while some are recognizable as software stacks, others are fuzzier, more heterarchical than hierarchical. ¹³ Beyond software, is the generic composability of any one layer in relation to another within a generative vertical platform that may help qualify systems as stacks.

52

Platform and Stack, Model and Machine

Stack architectures are also conceptual strategies for design, not just for description, and they are not only conceptual architectures, they are models for actual architecture as well. Le Corbusier ’ s Five Points toward a New Architecture is a strong stack, as embodied in Villa Sovoye and the vertical platform for five essential but undetermined programs. ¹⁴ The building may be “ a machine for living in, ” but the Five Points stack is the machine for making machines. Constant ’ s ever-changing New Babylon speculative urban system was redesigned again and again over the span from Sputnik to the OPEC embargo. It changed shape constantly, but one durable characteristic was the notion of a new city designed on top of the old one in two exclusive stacked layers. It imagined the new city as a landscape of vast multilayered networks and as continuous territories of ludic interfaces and opportunities, defined not by relation to a master ground plane but to the horizontal and oblique vectors of movement up and down the exploded sectional program. It was to be based not on functional regulation but on the feedback systems of play and serendipitous interaction. This project in turn inspired Rem Koolhaas ’ s revision and expansion of Mies van der Rohe ’ s sectional diagram into a generic principle of scale, for which the vertical juxtaposition of unlike programs in a single structure allows them to interoperate with as much mutual transparency or opacity as might be required, or which could be staged for optimizing spatial performance. This is seen perhaps most dramatically in the horizontal skyscraper OMA ’ s (Office of Metropolitan Architecture) 1972 conceptual project, Exodus: Voluntary Prisoners of Architecture, in which residents pass from layer to layer as they move through the discrete biopolitical stages of their lives. ¹⁵ Other architectural stacks are even more graphically explicit, such as Gordon Matta-Clark ’ s slices through stories of buildings, Robert Smithson ’ s concentric-layered world maps, and the stratified landscapes of MVRDV ’ s Hannover 2000 exposition pavilion that stacked and segmented artificial nature and program into a hyperdense world-in-a-box. Elsewhere, stack perspectives erupt uninvited and unintended. The verticality of flattened systems is seemingly uncontainable. While world maps render space in x- and y-axes, no linear geometry without thick verticality could represent the most entrenched geopolitical conflicts, rational and irrational alike. Consider Israeli architect Eyal Weizman ’ s multidimensional maps of the overlapping and interweaving claims of sovereignty in Israel-Palestine, showing that no horizontal cartographic linear delineation, or any regular vertical elevation all by itself, can finally describe, let alone govern, the multidimensional violence of that particular jurisdiction-intensive politico-theological matrix. Multiplications of the plane and rotations of perspective that move the flattened into the vertical are prevailing. Everywhere are stacks, good ones and bad ones, big ones and little ones, and many of them agglomerating into larger and larger platforms.

Platform and Stack, Model and Machine

53

The architecture of The Stack, this one particular megastructure of planetary-scale computation, is an interoperable physical-informational system of systems, distributed under, onto, and over the surface of the globe, with its layers organized into a patchy, uneven vertical section. As said, The Stack is composed of geologic, humanistic, and mineral layers charging feedback loops between these. As a cybernetic landscape, The Stack composes both for equilibrium and for emergence, one oscillating into the other for diagonal purposes in barely accountable rhythms. The state conditions (and literally for governance, the condition of the States that its platform logics describe in advance) are derived both from stacks as abstract diagrams and, through its unenumerated operations, as real existing machines. In turn the infrastructural sovereignties of The Stack may, in principle, emerge from either of these. It can be derived from its career as accidental megastructure, which itself may or may not be the model for geogovernance to come, or from its immediate, projective, and potential designability. It goes both ways. Today, The Stack that we can analyze frames the one we can conceive, just as the one we can conceive frames the one we are beginning to realize. Alternatives are conveyed from its distortions.

The Stack ’ s disciplining of communication as an ecology of isomorphic techniques makes the world appear as a system that demands from us a constant redesign of its ever more granular interoperations. The history of these technologies is also then the history of multiple competing communication standards. Protocol politics is always rough trade because to control the standard is to influence the economies it enables, which is to influence how they interrelate with other systems and the meta-economies those interoperations in turn give rise to. As should be plain from current events, the interweaving of otherwise incommunicative hard and soft systems into new assemblages continues apace, and so the politics of standards (e.g., open source, intellectual property, net neutrality, encryption) becomes integral to the “ democracy ” of infrastructure and to the little sovereignties of everyday life. Looking back, it is not coincidental then that formal systems theory and information theory appear historically concurrently and are part of the larger crest of cybernetics. The discernment of information as a first-order principle of material difference in the twentieth century would come to all but consume the very definition of systems tout court. The study of information bridged linguistics, symbolic logic, biology, chemistry, art, literature, and the theory of calculus with the practical engineering problems of automated logarithms, algorithms, cryptography, and long-distance signal transmission relay. In turn, the modeling of all of these and more as forms of information, as well as the conception of distributed multimodel apparatuses ultimately as information systems, becomes an overriding epistemological ambition of twenty-first-century globalization. This registration of all systems as information systems a priori tracks software ’ s migration from military logistics to consumer footprints. In this enforced translation of any thing into the status of information within a system, all things may possess their worlds and be possessed by their worlds only to the extent that they possess the attributes necessary for intermodal communication with other platform systems. Whether for bits or atoms, numbers or nectarines, no impedance mismatch can disallow the activation of that intermodality, and so compatibility within a given scale as well as the interoperability between scales, becomes itself the critical vernacular definition of computability as an economic technology. As all systems come to mean information systems, then computation, which otherwise might be defined differently, comes to refer to “ algorithms holding systems of information together. ” The Stack, as a particular megastructure, emerges from this history of systems conceived in relation to computation, and computation in relation to systems. It has inherited some of its limitations, ambitions, accomplishments, and blind spots and has evolved beyond others.

54

Platform and Stack, Model and Machine

1. Stack as Political Machine

The emergence of planetary computation as a global and intelligent system can be traced in broad strokes from perhaps Roman and Chinese military accounting, to the first Victorian calculators through to today, and it is marked by celebrated breakthroughs as well as long-ignored dead ends, some of which are eventually celebrated retroactively. By their appearances on the scene, it seems that every globalizing communication network, from printed books to telegraphy, railroads, radio, telephony, and television, was celebrated (and lamented) as the coming of some universal political community, messianic or degenerate or both. In their formative years, new regimes of digital global media are as well invested and inflated with world-historical importance, as signaling the ultimate arrival of a too long postponed cosmopolitanism. (Both Hegel ’ s political time and Kant ’ s political space were themselves conceived in intimate proximity of that most modern of global mass media, the network body of the state, which would continue to reinvent its own anatomy in relation to new media regimes.) If we are more used to living so much of our lives inside the shells of planetary-scale computational networks, we also observe that the political realities of universal global information turn out to be far more problematic, more mundane, and unusual than envisioned, feared, and hoped for. This disappointment itself may be the most timely message of the medium, but not necessarily its most lasting. Perhaps the persistent utopianism around communications infrastructure still works, not because of how well it predicts the outcome of large-scale technological interventions, but because as the genesis of productive accidents, it is able to make room for otherwise unauthorized political and social forms in its wake. The Stack, like any other technology of such scale and significance, both constitutes a new political-geographic order and enforces an existing cultural-economic order already in place. ¹⁶ It does each in different ways and at different locations, and the untangling of these is part of the design brief. Toward this, we have to do more than map platforms; we have to learn to read them and interpret them.

Platform and Stack, Model and Machine

55

Conway ’ s law, coined in 1968 by programmer Melvin Conway, states that “ organizations which design systems ... are constrained to produce designs which are copies of the communication structures of these organizations. ” Put differently, “ in order for two separate software modules to interface correctly, the designers and implementers of each module must communicate with each other. Therefore, the interface structure of a software system will reflect the social structure of the organization(s) that produced it. ” A corollary law might demonstrate that over time, the inverse is equally true: organizations come to take on the characteristics of their interfaces. If it goes both ways, then homologies between organizer and organized make the detection of cause and effect between cultural and technical systems rather difficult. For example, and to extend this problem to the largest scale, an understanding of the ongoing emergence of planetary scale computation cannot only be understood as a secondary technological expression of capitalist economics. ¹⁷ The economic history of the second half of the twentieth century is largely unthinkable without computational infrastructure and superstructure. In accounting for that transformation, it is not at all clear whether the computational technologies are more or less foundational than the economics that organized them and that they organize (even assuming that we could analytically separate the two, so as to put one in the fore and the other in back). Instead of locating global computation as a manifestation of an economic condition (as both its means of production and it superstructural expression), the inverse may be equally valid. From this perspective, so much of what is referred to as neoliberalism are interlocking political-economic conditions within the encompassing armature of planetary computation. The entwined polar positions of Sunnyvale, Caracas, Beijing, Brussels, Tribeca, and Tel Aviv don ’ t integrate capital and resource markets into network societies on their own, but are themselves “ computed ” into these arrangements. Either way, it is possible to delineate structural causality between technological and socioeconomic systems only in mode abstractions, because one always contains the other and is contained by the other at once. We cannot, for example, finally locate computation technologies as a base and information culture as a superstructure, bound together either through capitalintensive modes of production and exchange or through computational flows directly determining systems of valuation and exchange in their image. Rather, we could do so, but only with abstractions that are easily turned inside out. Neither serves as the last instance of the other, though today, neither can be defined without the other. Instead, we should work with this mesh on its own less mechanistic terms. Jameson, fo example, offers an injunction to treat such arrangements as active temporal operations rather than as a fixed architecture, “ basing-and-superstructuring. ” ¹⁸ We may think of foundation as a verb, base-ing , and to seek out how specific material technical systems come to take on causal force and when the same do not. Such flexibility might allow us to differentiate, for example, when the discursive structure of the relational database drives not only the information access policies of a company or state, but also in turn the form of its organizational hierarchies, and when the inverse is predominantly true, such as when the laws and logistics of trade channels structure the form and content of interoperable supply chain management software and the database designs on which it depends. In locating The Stack within the intercourses of economics, culture, and technology, both Conway ’ s law (that organizations design systems in their image) and our inverse Conway ’ s law (that systems and their interfaces produce organizations in their image) are interpretive tools that are useful to keep at hand.

56

Platform and Stack, Model and Machine

As a platform to be read and interpreted, The Stack clearly sits on both sides of this coupling of culture and technology. It relies on software as both a kind of language and a kind of technology, of algorithms of expression and the expression of algorithms, and this twisting of the conceptual and the machinic can sometimes bring emotional distress. ¹⁹ For some, an apparently universal convertibility of social systems into software systems motivates euphoric convictions in the instantaneous self-realization of networked individuals, a particularly Californian enthusiasm spanning from the ingenious to the idiotic. (The so-called California Ideology is not what I am referring to here. That term was always a simplistic New Left chestnut that crudely lumped Survival Research Laboratories and Page Mill Road venture capitalists into one cohort.) ²⁰ However, the extremities of convictions also give way to a more nuanced complex of platforms that not only augment force and authority but constitute first-order modes of authority on their own (they are discussed in following chapters). These are the geographic powers to be further decided and designed, or left to go stale and rot on their own. For example, the ability of some platforms to absorb and recognize patterns in end User behavior might mimic how markets resolve fluctuations of price, but its formal centralization also allows for higher-level forms of planning, investment, and equity that states are, ostensibly, steering on their own. Contemporary Cloud platforms are derived from more specific systems of user-facing interfaces and services (discussed at length in the Cloud chapter). The intelligence of User interactions provides core content that is aggregated, optimized, and made more visible, more immediate, more standardized, more interoperable, more mobile, and therefore more valuable both to Users and to the platform than it would be otherwise. How far can that go?

For some, the capacity for platforms to operate in this way suggests striking similarities with the hopes of socialist planners to engineer a pricing and planning mechanism that could observe, analyze, calculate, produce, and distribute materials and goods according to principles of rational evaluation instead of the anarchic vagaries of supply and demand. ²¹ Francis Spufford ’ s 2010 historical novel, Red Plenty, did much to respark interest in this overlooked period in the history of political computer science, and in Khrushchev- and Brezhnev-era Soviet economists and cyberneticians in particular. ²² At that time, planners and programmers had access to what is by today ’ s standards minuscule computing capacity to calculate patterns, pathways, and contingencies, but contemporary supercomputing systems not only could orchestrate and optimize the pricing and dissemination requirements of large economies; they do it every day. ²³ Those planners and programmers also labored under centralized authoritarianism, and so for our efforts to plot out where else platform economies can be made to go and what alternatives to Anthropocenic economics are possible, it is not suggested that we look back on midcentury regimes for all the key clues. However, the clear homologies between the aspirations of Soviet cybernetics and the accomplishments of Google, for example, to model and govern superpower-scale digital economies, and the genealogies that link the latter to the former, at least testify against the notion of an intrinsic bond between capitalism and computational megaplatforms. We may anticipate that to some significant extent, the dovetailing of the future evolution of both agendas will transform one another and may even allow one to fully envelop the other: neither state as machine nor market as machine because the platform is state, market, and machine at once. Some Marxian articles of faith (such that once global technological means of production and valuation have reached some threshold level of efficiency and ubiquity, such that continuance of management by capital is not needed, then things will give way to a self-regulating infrastructural commonwealth) may have surprising interpretive value for the next century even if it works out in ways utterly different than originally and normally conceived. As many on the left and the right have postulated, the acceleration of capital flows through computational megaplatforms such as these may, in the long run, do as much to undermine the modern function of exchangeable property as it does to radicalize it (and perhaps the former because of the latter). We will have to wait and see what will and will not “ wither away ” should planetary-scale computation approach peak platform optimization and ubiquity, but in the meantime, we have other historical examples of proto-Stack governments to consider and to interpret.

Platform and Stack, Model and Machine

57

1. Stacks That Were and Might Have Been

In 1970, British cybernetician Stafford Beer was commissioned by Unidad Popular, the new socialist Chilean government of Salvador Allende, to design the platform for a new computer-controlled economy, a project that came to be known as Project Cybersin (the name is a conjunction of “ cybernetics ” and “ synergy ” ). ²⁴ The proposed network would have organized the entire Chilean economy according to, among others techniques, a twelve-layer concentric platform model, running from the worker himself (center layer), to successive layers of the crew, workshop, department, firm, line, sector, branch, industry, state economy, central government, and finally enveloped by the twelfth, and final, layer of the whole nation. As seen in Beer ’ s diagrams, layers in the system could recursively influence layers it surrounded, with any one factory floor or shipping port location sending status information regularly into the platform by a network of telex machines. This feedback loop seems like rocks and sticks compared to today ’ s multivariate stochastic logistical lattices, but in the early 1970s, it was practically science fiction. Neither that Cybersin network nor its planned control room straight from Captain Kirk ’ s bachelor pad survived the Pinochet coup. The Chilean army and its sponsors apparently didn ’ t see much value in flat, decentralized economic ecologies sutured by cutting-edge information networks. Hmm. ²⁵ Beer ’ s “ stack ” was based on his viable system model, “ a five-tier structure based on the human nervous system, which Beer believed existed in all stable organizations — biological, mechanical, and social. ” Biological system metaphors map onto modern social theory in diverse ways and while some emphasize equilibrium, others emphasize emergence. ²⁶ For Beer ’ s Project Cybersin, it was the latter. The aspiration of the platform was to constitute and compose a systemic state condition, literally a socialist nation-state condition, and bring it into being. The platform sovereignty attempted by his stack was generative ex nihilo. Unlike the deeply centralized planning mechanisms of the Soviet Gosplan, the reporting, planning, and coordination of Cybersin ’ s architecture was meant to be decentralized and democratic. Beer himself was less interested in Marxian theory of history than in the revolutionary potential of autopoietic cybernetics as a form of governance, and the presumed effect of information flows to make systems less hierarchical, more composable, more vital and durable.

58

Platform and Stack, Model and Machine

Meanwhile in Japan, a platform of cybernetic equilibrium has been in continuous development since 1984, one for which the normative apparatus of the nation and its interpolation of objects and subjects within its industrial economy is presumed, reified, served, and conserved. Ken Sakamura ’ s TRON (an acronym for “ The Real-Time Operating system Nucleus ” ) is the basis for a “ computer everywhere ” infrastructure that envisions a ubiquitous national computer network built on a distributed real-time operating system among a vast network of objects and terminals of multiple scales and complexities, a blend of an Internet operating system and Internet of Things communication formats with ubiquitous ID systems. ²⁷ In its heyday, TRON was an architecture and interfacial framework by which parts of the whole national industrial system, from cell phones to cars to factories to municipal infrastructure, could communicate and be addressed by similar and related software frameworks. Those frames were forked and developed as different subarchitectures, each incorporating a different subset of an overarching Japanese computational ecology: ITRON, JTRON, BTRON, CTRON, MTRON, and STRON for mainframes, industrial uses, telecommunications, cell phones, end user terminals, and on and on. The industrial variant, ITRON, for example, is widely used in Japanese embedded systems. It was designed to be what Sakamura called “ open-open ”

Platform and Stack, Model and Machine

59

in the communication between parts and components and freely available to be reprogrammed for specific implementations. Taken as a whole, TRON was considered at one point “ the most popular operating system in the world. ” ²⁸ Its success, however, has been limited by the insularity of the Japanese technical ecology, and the contiguity, coherency, and delimitation of the Japanese state system. ²⁹ The ultimate boundary of its walled garden would prove to be as unambiguous as an island ’ s border, and so the growth of TRON was both enabled and curtailed by Japan ’ s Garapagosu-ka or “ Galapagos syndrome. ”

Sakamura ’ s stack was constituent and curatorial; Beer ’ s was constitutive and generative. Beer ’ s model posited a nested series of socioeconomic scales, from worker to nation, through which regulated information would be reported, analyzed, and governed. Sakamura ’ s model distributes operations among widely dispersed components sharing data directly or indirectly for separate uses (e.g., industrial, civic, interpersonal) and so lubricating intermodal communication between people and people, people and things, and things and things. Beer ’ s and Sakamura ’ s visions are asymptotic. Both sought to design a platform infrastructure that would integrate a national society by integrating its material economies into a master computational system, but each is animated by a different conception of that task. Beer ’ s assignment was to help engineer a new nation into being through cybernetics, and so the key diagrams of his stack depict the socioeconomic scales that would come to participate through that system. For Sakamura ’ s Japan, the program of his stack is to intensify a national and cultural equilibrium already established, and so his images depict not a new social order (as Japan ’ s organic stability could be presumed) but the technical network layers that would be made to serve it. Beer ’ s diagram was of the macrosocial emergent effect of platform sovereignties, and Sakamura ’ s was the inverse, a technical harmonization of a social foundation. The constitutive design imagined the social and the cultural as an effect of the technical intervention, and the constituent design imagined the technical as a function of the social and cultural, and so here, both sides of software as language and software as technology dichotomy are exemplified and mirrored. For The Stack, the essential forces of the generative and the regulatory, equilibrium and emergence, constitutive and constituent force, remain equally foundational for one another. The Stack freezes, radicalizes, and reinforces models of governance and macroeconomics at the same time as it dismantles them, builds geographies above and below them, and undermines their ability to reproduce themselves. Sometimes it accomplishes one by doing the other.

The ongoing design of The Stack is for an architecture that is equally technical and conceptual, drawn by both its diffuse instrumentality and its physicalized abstractions. It is well suited to reflect even politico-theological aspirations and can easily synthesize an ideal liturgy of preferred signals and echo them back at specific Users (as discussed in the Interfaces chapter), even as its ability to absorb and revalue new content (toward inflation or deflation) is programmed to be agnostic and omnivorous as strategy dictates. As we conceive possible futures for The Stack platform, in what way might the generative aspiration of Cybersin do more work for more Users than the curatorial aspiration of TRON, or vice versa? As ideal types, TRON is designed to optimize coherent divisions, whereas Cybersin introduces communication across the boundaries of scale. Cybersin focuses on the flow and valuation of goods and actions taken, and TRON on the embedded intelligence of infrastructure and equipment. Cybersin seeks to enroll all actors in the economy into its ongoing planning and evolution, while TRON seeks to allow steering authority a more transparent chain of command. Beer and Sakamura would likely agree, however, that any platform architecture will succeed not through the premeditated ingenuity of its original schemes that will always prove too brittle, but through how it is taught to accommodate and validate unforeseeable new programs, and to do so as simply as possible.

60

Platform and Stack, Model and Machine

1. The Stack We Have

The Internet is built out on stacks not so dissimilar to these. Its eminence and its monotony, both available in excess, are functions of the regularity and resiliency of several specific “ stacks, ” variously abstract models and real technical machines. As its stacks are far more pervasive and powerful than Cybersin and TRON, their dissection demands more reading and interpretation. They are the result of work by well-known scientists and engineers (more on them below), anonymous ingenious workarounds, and coordinated tactical responses of established industrial Users. Equal measures of emergence and equilibrium are at work in the interplay of anticipatory design and realtime problem solving. The shape of The Stack, our accidental megastructure (which contains this Internet but also much more), is also built on some of these models and their particular governing steerage of information networks. The history of Internet stacks also makes clear that the authoring of architectures can produce metasystemic accidents that in turn can themselves counterauthor and redesign the platform in unintended but successful ways.

Consider the lessons of the OSI (open systems interconnection) network model and the TCP/IP network model, on which global Internet connectivity is now largely based. The specification of both standards began in the 1970s, and the latter was more fully adopted by the mid-1980s. For OSI, the network represented has seven discrete layers, from the application layer addressing Users at the “ top ” down to the physical layer (which today may be fiber-optic cable channeling pulses of light). Below the application layer are, in descending sequence, the presentation layer, the session layer, the transport layer, the network layer, the data link layer, and then the physical layer. In simple terms, a message is sent by a user down the stack, layer by layer, until it is transmitted laterally across the physical layer to the receiving node across the street or across the ocean. Upon being received, the message travels back up the stack, from the physical layer to the application layer, and is read by the next user. Each of the layers gathers, specifies, and links heterogeneous technologies together into a functional stratum. ³⁰ On each layer, an instance provides services to and requests service from the layers above and below, and can also pass information laterally to its corresponding layer (i.e., network layer to network layer, as in many so-called level 2 networks linking financial traders and trading centers). Key to the success of this modular model is its flexibility in absorbing future technological innovations that can be introduced at any given layer (e.g., fiber optics instead of copper wire at the physical layer, better router software, an application with better features and security) without disrupting the existing components, so long as the new technology adheres to the protocols established by the platform model that allow it to communicate with its vertically adjacent layers, above and below. In principle any machine could be inserted in a layer of the network if it can adhere to the necessary grammar that would allow it to communicate with its most proximate neighbors. For The Stack, the OSI model serves as a literal and technical prototype for how network architectures operate between very small and large scales and, as the primary abstraction, or universal diagram, for how its heterogeneous participants can arrange communication in a vertical assemblage, now at a megastructural scale.

Platform and Stack, Model and Machine

61