Cloudware - Cloud Computing Without Compromise

The Web has succeeded by allowing individuals and companies to add their own unique capabilities to The Net. Cloud Computing should develop the same way. Developers shouldn't be forced to compromise on the selection of development platforms, software, or security. Cloudware is an architecture intended to provide an open framework allowing the development of a cloud computing environment that's rigorous enough to take on any web or enterprise application.

Cloud Computing without compromise - cloudware

Cloudware incorporates the fundamental building blocks used in developing today's most popular applications; storage and computing, software catalog, definition and control, plus how they all relate to each other. More importantly, the architecture is vendor agnostic so that third party vendors, not just 3tera, can participate in the system. The Cloudware architecture supports the most popular operating systems - Linux, Solaris and Windows - and is targeted toward clients who want to explore the extreme scale and flexibility of cloud computing infrastructures quickly and easily.

Cloudware will be rolled out in stages over the next twelve to twenty-four months, but much of the technology necessary to implement Cloudware has been proven in our AppLogic™ grid operating system over the past two years. Some of the upcoming features include adding support for Solaris and Windows to existing support for Linux; choice of multiple data centers worldwide, pre-built MySQL clusters, database replication appliances and NAS integration with third party storage solutions

When virtual appliances aren't running, they consume no processing resources and only a small amount of storage. As such, rather than using them sparingly as traditional software is, virtual appliances can be packaged with every application that uses them. In essence, the virtual appliances form a disposable infrastructure on which the application relies while operating. When the application is run, the virtual infrastructure it requires is created dynamically on the grid, maintained while it runs, and disposed of when it stops.

The use of virtual appliances makes applications scalable within AppLogic as well. Each virtual appliance has a range of resource usage over which it produces valuable output. The actual resources used in production are only assigned at runtime. For instance, a web server virtual appliance may require 5% of a CPU and 64MB of RAM to respond to a query and may stop providing greater output at 50% of a CPU and 1GB of memory. A two-tier application that includes this virtual appliance in addition to five or six other virtual appliances, may run with as little as 25% of a CPU and 300MB RAM, but still be able to scale to more than 3 servers and 6GB of memory.

^ Top of Page


Cloudware Players

Subscribers

Small and medium businesses, Web 2.0 and SaaS companies, as well as enterprises – anyone who needs to stand up IT infrastructure and have working web applications and services.
Subscribers can have their applications operate in the cloud, without the need to own or manage servers, data centers, network peering, etc. They can deploy any desired architecture, middleware, including existing applications; can scale applications per their needs, and operate them anywhere in the world, paying only for what they use.


Data Center Operator

Data center operators and hosting providers "publish" computing resources – such as servers, storage and network connectivity – making them available to subscribers. Data center operators include hosting providers, managed service providers, enterprise datacenters and other clouds. The data center operators determine the prices for the resources they publish and who can use them – from individual subscribers (e.g., when an enterprise data center adds private resources for use by other subscribers in the enterprise), to general use by any subscriber. In addition, data centers can publish their excess capacity, or have the unused servers shutdown to conserve power until needed.


Publishers

Independent software vendors, virtual appliance vendors, infrastructure, platform and tool vendors can publish appliances, ready-made architectures, and whole ready-to-run applications in the global catalog. Publishers determine which subscribers have access to what published resources and at what price. Virtual appliances allow, among other things, all hardware appliance vendors to provide software equivalent of their appliances, including firewalls, load balancers, security appliances, etc.; platform and middleware vendors can provide ready-to-use packages of their software that can be used without complex installation and configuration. IT professionals can productize their expertise by publishing ready to use architectures: LAMP, Ruby-on-rails, J2EE, including scalable versions, such as clustered database servers, application servers, etc. Verticals can publish their applications in a ready-to-run form that can be delivered by managed service providers or used by customers directly.


Vendors for Integrated Web Services

Vendors can provide value-adding web services that are available to all subscribers.
Examples include advanced monitoring tools, billing services, transaction monitors, lifecycle management, storage and policy engines, etc.


Providers for Outsourced Services

Outsourcing providers can publish their services and make them easily available on the cloud, whether the services are application development, monitoring, support, or application management

^ Top of Page


Cloudware Elements

Resource Pools

The resource pools provide raw compute power, storage and network connectivity to applications running in the cloud. Resource pools can be located anywhere in the world; even individual datacenter operators can publish multiple classes of resource pools – in terms of network connectivity, power, services, etc. Commodity servers are turned into resource pools by installing the AppLogic execution engine; other resources – such as 3rd party clouds – can also be published by providing the appropriate interfaces. The resource pools are controlled by the Infrastructure Delivery Network via web services interfaces.


Global Catalog

The global catalog is a worldwide distributed catalog service where various publishers make their appliances, architectures and applications available to subscribers. Multiple catalogs can be managed and accessed by subscribers, allowing software publishers to organize their catalogs, and customize them for their target markets. The global catalog includes a versioning and distributed caching system that makes catalogs available to any application anywhere in the world.


Control Interface

The control interface includes a set of user interfaces and APIs for controlling applications and services running in the cloud. It includes subscriber portals, dashboards, monitoring screens, infrastructure design tools, development tools (e.g., Eclipse plug-ins) and command-and-control web-based interfaces. In addition, the control interface includes web services APIs for “programming” the cloud.


Infrastructure Delivery Network

The infrastructure delivery network aggregates the different components of the cloud and their separate instances in a cohesive, distributed cloud. It provides authentication, access controls, registration of resource pools, control interfaces and catalogs; it deploys infrastructure from the catalogs to the resource pools as necessary to provide the services specified through the control interface; provides data source for the integrated web services; facilitates the interactions between components of the cloud; manages complex transactions during deployment and migration. The infrastructure delivery network is a distributed service, providing resilient, highly-available and localized nervous system for the cloud.

^ Top of Page


Cloudware Clients

Various users on the Internet, whether they are connected with laptops, desktops, or mobile phones. Services and applications running in the cloud are accessed over existing protocols and are indistinguishable from services running on traditional architectures, except for their improved scalability, availability, etc.

^ Top of Page