Digital Ecosystems in the Clouds: Towards Community Cloud Computing

I. INTRODUCTION

• Cloud Computing's value for organisations outsourcing ICT infrastructure is accompanied by concerns about vendor control, information privacy, and the growing carbon footprint of data centres.The paper links exponentially growing data centres to environmental concerns.Source: §I · p. 1
• Digital Ecosystem principles and community ownership can address these concerns by using spare resources of networked personal computers to form a virtual data centre.The proposed Community Cloud is a socio-technical conceptualisation for sustainable distributed computing.Source: §I · p. 1

II. CLOUD COMPUTING

• In the typical configuration, a central Cloud housed in one or more data centres serves application consumers, with resource provision and coordination centrally controlled.Figure 1 marks consumption in green, provision in yellow, and coordination in red.Source: Figure 1 · p. 1, §II · p. 1
• Cloud Computing provides dynamically scalable, often virtualised resources as Internet services, while users lack knowledge of, expertise in, or control over the supporting infrastructure.It is described as a commercial evolution of academia-oriented Grid Computing.Source: §II · p. 1

A. Layers of Abstraction

• Cloud offerings fall into different abstraction levels aimed at different market segments, although their qualifications and interrelations are often unclear.The paper presents this categorisation as a way to clarify the offerings.Source: A. Layers of Abstraction · p. 2

1) Infrastructure as a Service (IaaS) [18]:

• IaaS providers such as Amazon and Mosso offer machine instances that developers operate and scale manually when performance limits are reached.This level gives developers responsibility for running arbitrary software with few development-method compromises.Source: 1) Infrastructure as a Service (IaaS)… · p. 2

2) Platform as a Service (PaaS) [21]:

• PaaS environments such as Google App Engine abstract machine instances and allocation, while imposing application-design constraints such as using key-value stores instead of relational databases.Source: 2) Platform as a Service (PaaS) [21] · p. 2

3) Software as a Service (SaaS) [15]:

• SaaS provides consumer-facing applications with online resources and storage, including email, office suites, and business software.It differs from traditional websites or web applications by interfacing with user information such as documents.Source: 3) Software as a Service (SaaS) [15] · p. 2
• Cloud actors may take multiple roles, but within a Cloud a single vendor occupies the provider and controller role.Application developers build services for end users using resources provided by the vendor.Source: 3) Software as a Service (SaaS) [15] · p. 2

B. Concerns

• The paper frames concerns about the Cloud Computing model as a central issue for the following discussion.Source: B. Concerns · p. 2

1) Economics of Failure:

• Vendor-driven monocultures can turn outages into cascading, system-wide failures, placing Internet resilience at risk despite the uptime advantage of Cloud solutions.The Amazon S3 outage is cited as taking down dependent businesses.Source: 1) Economics of Failure · p. 2

2) Convenience vs Control:

• Cloud convenience brings vendor control and privacy concerns, while providers may face a conflict of interest when serving SMEs and start-ups that could challenge their consumer-facing businesses.Vendors have access to resources stored on their Clouds.Source: 2) Convenience vs Control · pp. 2–3

3) Environmental Impact:

• Data-centre growth increases carbon-footprint concerns, and virtualisation to maximise resource utilisation has not resolved the problem.The paper also notes legislative, power-grid, and financial incentives to address it.Source: 3) Environmental Impact · p. 3

2) Convenience vs Control:

• A meta-Cloud can avoid vendor lock-in only to an extent, while Eucalyptus provides privacy and avoids lock-in only for organisations with their own data centres.The paper says this private-cloud arrangement does not meet its definition of Cloud Computing, which avoids requiring one's own data centre.Source: 3) Environmental Impact · p. 3

III. DIGITAL ECOSYSTEMS

• Digital Ecosystems are distributed, adaptive socio-technical systems that aim to support sustainable regional development and help local actors form value networks.They extend local culture and identity into global interaction.Source: §III · p. 3
• Virtualisation creates a virtual version of a resource that can be stored, migrated, duplicated, and instantiated as needed.The passage links this flexibility to improved scalability and workload management.Source: §III · p. 3
• Network-based economies share access to services, but Digital Ecosystems still lack scalable resource provision and therefore need community-based infrastructure.The section argues that otherwise they risk relying on vendor Clouds at the infrastructure level.Source: §III · p. 3

IV. COMMUNITY CLOUD

• Community Cloud Computing would shape underutilised user-machine resources into a Cloud where nodes can consume, provide, and coordinate resources.Figure 3 uses green for consumption, yellow for provision, and red for coordination and administration.Source: §IV · p. 3, Figure 3 · p. 3

A. Conceptualisation

• The Community Cloud combines Cloud Computing, Digital Ecosystems, and Green Computing as a vendor-independent paradigm for community Clouds.The proposal removes dependence on vendors such as Google, Amazon, and Microsoft.Source: A. Conceptualisation · p. 4

1) Openness:

• By removing vendor dependence, the Community Cloud extends the open-versus-proprietary struggle into hosted services.The passage identifies code, standards, and data as earlier dimensions of that struggle.Source: 1) Openness · p. 4

2) Community:

• Community ownership makes the Community Cloud both a social structure and a technology paradigm, with economic scalability intended to sustain competition and innovation.The passage warns that without this scalability, competition could diminish and innovation could be stifled.Source: 2) Community · p. 4

3) Graceful Failures:

• Diverse supporting nodes are intended to avoid dependence on one organisation and limit system-wide cascade failures, with unaffected nodes compensating when others fail.The passage says occasional failures would be graceful, nondestructive, and involve minimal downtime.Source: 3) Graceful Failures · p. 4

4) Convenience and Control:

• Community ownership provides democratic distributed control, avoiding an inherent conflict between convenience and control.Source: 4) Convenience and Control · p. 4

5) Environmental Sustainability:

• Using underutilised user machines is expected to require less energy than dedicated vendor data centres and produce a significantly smaller carbon footprint.The Community Cloud is described as growing and shrinking symbiotically with community demand.Source: 5) Environmental Sustainability · p. 4

B. Architecture

• The architecture distributes server functionality across user-machine nodes, shaping their underutilised resources into a virtual data centre.The passage notes that implementation poses challenges at multiple levels.Source: B. Architecture · p. 4

1) Core Infrastructure:

• Isolated virtual machines would form a fully distributed peer-to-peer network supporting distributed identity and coordination.The virtual machines provide a sandbox for guest code and make host resources safely available to the Cloud in a measurable manner.Source: 1) Core Infrastructure · p. 4
• The network must avoid single points of control and failure, while distributed identity and trust track node history and multidimensional characteristics.Trust considerations include uptime, performance, and reputation.Source: 1) Core Infrastructure · p. 4

2) Resource layer:

• The resource layer would offer raw storage and computation at the PaaS layer, with service deployment providing the SaaS layer.Source: 2) Resource layer · p. 4
• Distributed computation would draw on Grid Computing to coordinate computational capacity offered by nodes.Source: 2) Resource layer · p. 4
• Storage would combine distributed storage, databases, and key-value stores to address differing data and availability needs, while encrypting information on remote nodes.The passage says information is unencrypted only when cached on the user's node.Source: 2) Resource layer · pp. 4–5
• The Community Cloud may require more bandwidth than vendor Clouds, so it would use peer-to-peer distribution approaches to use available bandwidth efficiently.The passage identifies sudden rises in demand as a risk to parts of the network.Source: 2) Resource layer · p. 5
• A community currency redeemable for community resources would reward users who contribute valued resources.Its prices could fluctuate with market demand rather than fixing relative resource costs in advance.Source: 2) Resource layer · p. 5

• Community Cloud replaces stand-alone services with services composed by default from resources contributed by multiple participants.Source: 3) Service Layer · p. 5
• The distributed repository must persist service pointers, semantic descriptions, and executable code, including when service-producing nodes are absent.Source: 3) Service Layer · p. 5
• When a suitable node lacks a needed service, it retrieves and instantiates a repository copy, supporting responsiveness to unpredictable traffic spikes.Source: 3) Service Layer · p. 5
• Service descriptions should state resource costs for pre-execution budgeting and post-execution community-currency payments; inaccurate costs can burden users or terminate services prematurely.Source: 3) Service Layer · p. 5
• Remote execution must be secured against compromised nodes that could access the executed services’ business logic.Source: 3) Service Layer · p. 5

• Wikipedia’s donation-dependent funding needs continual monetary contributions, while advertising has caused community conflict and risks public trust.Source: §V · p. 5
• A Community Cloud could offer Wikipedia scalable, self-sustaining resource provision compatible with its communal nature and without compromising content trust.Source: §V · p. 5
• Active participant nodes earn community currency by fulfilling requests, which users can spend to search Wikipedia or retrieve pages.Source: §V · p. 5
• Editing a page requires updating distributed storage through the node network, likely using eventual consistency.Source: §V · pp. 5–6
• The user-funded provision model also applies to for-profit businesses and creates a micropayment scheme that lowers entry barriers for innovative start-ups.Source: §V · p. 6

• Community Cloud is a socio-technical conceptualisation for sustainable distributed computing and an alternative to Cloud Computing, blending its usage scenarios with Digital Ecosystem principles.Source: §VI · p. 6
• The model uses spare resources from networked personal computers to provide data-centre facilities for a cloud platform’s community.Source: §VI · p. 6
• The authors hope Community Cloud will encourage vendor-Cloud innovation through creative tension analogous to that between open-source and proprietary software.Source: §VI · p. 6