Deployments Glossary
Paperspace Deployments are containers-as-a-service that allow you to run container images and serve machine learning models using a high-performance, low-latency service with a RESTful API.
This glossary defines the core concepts behind Paperspace Deployments to help build your mental model of how deployments work and understand what the documentation is referring to when it uses certain terminology.
Continuous integration and continuous delivery/deployment (CI/CD) encompass several automated processes commonly performed in software development, such as building, testing, and deploying a software application. CI/CD reduces human intervention during the maintenance processes of software applications.
A container is a single running instance of a service, worker, or static site resource that you have shipped on App Platform, and is in essence a live, runtime instance of an image (the artifact produced by the build process).
A central processing unit, also called a processor or CPU, is an essential piece of computer hardware that executes the instructions from a computer program.
In a denial-of-service (DoS) attack, a threat actor renders information systems, devices, or other network resources inaccessible by overwhelming the target with malicious traffic. A distributed denial-of-service (DDoS) attack is a type of DoS attack where the overloading traffic originates from multiple attacking machines, amplifying the severity of the attack.
An endpoint is a specific URL or network address that a client device or application uses to interact with a server or service. Most internet connected apps and websites communicate with endpoints to retrieve information, process user interactions, and perform other processes. For example, when you open a weather app on your phone, the weather app requests for the latest forecast information from a server endpoint.
An environment variable is a in-memory key/value pair that developers use to store configuration details about their app.
Ephemeral files are temporary files used for program execution and tasks.
An epoch is a single iteration through a dataset when training a machine learning model.
Gradient deployment allows machine learning models as a model as a service where the model is placed in a special container where the model performs the same way every time it is used.
A Gradient model is a scalable trained machine learning or deep learning model.
Gradient volume is a special storage for storing datasets, models, and project files. It allows users to keep track of data changes over time as well as saves backups of the data.
Graphics Processing Unit (GPU) is a specialized processor meant to accelerate graphics rendering or processing complex mathematical calculations such as during machine learning training.
A handshake is the authentication process between two networked devices to ensure that both devices are who they claim to be.
A health check is a scheduled HTTP or TCP request that you can configure to run on a repeating basis to ensure that a service is healthy.
A hypervisor is a virtual machine monitor (VMM) that allows multiple operating systems to run simultaneously on one host computer.
The Internet Protocol (IP) is a communications protocol used to connect computers across a network, specifically the Internet. IP consists of rules and regulations for transmission of packets across a network including routing and addressing. IP ensures that the packets of data that travel across a network arrives at the correct location.
A key, or bucket key, is a unique data key for objects in a bucket.
Machine type is the set of virtualized hardware resources used for a virtual machine (VM) instance.
A network is a collection interconnected servers or computers that communicate and share resources amongst each other.
A notebook is a virtual environment where code is sequentially written and executed.
An Operating System Template or Machine is a disk image of a pre-designed or pre-configured version of a machine.
Random-Access Memory, or RAM, is a memory storage made to store and access memory on a short-term basis.
A runtime environment is an execution environment for deploying and running applications. Runtime environments contain the necessary applications, dependencies, and operating system to run an application. For example, when you create an app on App Platform, App Platform inspects the code and app resources, and then selects the appropriate runtime environment, such as Node, to run the app in.
S3-Compatible Object Storage, or S3 is a storage solution that uses the S3 API.
Software Development Kit (SDK) is a set of software tools and programs designed for a specific service or system.
Symmetric ciphers use the same key for encrypting and decrypting data.
Two-Factor Authentication (2FA) is a security method where the user uses two different authentication methods to verify their identity.
In machine learning, a workflow is a set of processes working towards a certain objective.