Anda di halaman 1dari 22

White Paper: Media Clouds

White Paper

Media clouds: How can cloud computing help in the content creation process?
Dening a cloud computing stack of applications, services and processes used for digital newsrooms

Authors: Mirko Lorenz, Linda-Rath Wiggins (Deutsche Welle) Paola Sunna, Alberto Messina (RAI)

The research leading to these results is partially supported by the European Communitys Seventh Framework Programme ([FP7/2001-2013]) under grant agreement n 257019. More Information: VISION Cloud Website

Note:
This whitepaper takes a look at different cloud computing models and actual offerings that are already on the market, from the perspetive of media companies. The content of this paper represents steps of our process to contribute to innovative future solutions. Partially the ndings where used for ofcial reporting for the EU. Still, we hope that this long version might be of value to other media companies or other researchers tackling the same questions, which is why we published this early research work on the website. Please note that this is work in progress.

White Paper: Media Clouds

Table of Contents
1. 2. Executive Summary!...................................................................................................3 Cloud Computing and Media Organizations!...........................................................5 2.1. Introduction ....................................................................................................5 2.2. External Challenges .......................................................................................5 2.3. Internal Challenges .........................................................................................6 2.4. Enhancing the digital supply chain with the media cloud ..............................7 2.5. Evolving new models for media .....................................................................8 2.6. Cloud Architectures Relevant for Media Companies .....................................9 2.6.1. Enterprise to Cloud!...........................................................................9 2.6.2. Enterprise to Cloud to End User!.....................................................10 2.6.3. Enterprise to Cloud to Enterprise!...................................................11 2.6.4. Private Cloud!..................................................................................11 2.6.5. Which cloud model for media companies?! .....................................11 Role of VISION Cloud!...............................................................................................13 3.1. About VISION Cloud ....................................................................................13 Market Analysis: What are the best solutions so far?!..........................................14 4.1. Basic model: A cloud-based media workow .............................................14 4.2. Workow: Teams ..........................................................................................16 4.3. Ingestion .......................................................................................................16 4.4. Filter, transform & search .............................................................................17 4.5. Production ....................................................................................................18 4.6. Publishing .....................................................................................................19 4.7. Distribution ...................................................................................................19 4.8. Popularity aka Usage Analytics ....................................................................20 Early conclusions: Media Clouds - what is there and what is missing?! .............22

3. 4.

5.

White Paper: Media Clouds

1. Executive Summary
This market analysis examines how cloud computing platforms could be used in media companies, taking into account basic, theoretical models (e.g. Software as a Service, etc.) and real-world platforms, that perform particularly well to accomplish tasks in the media production process. Finally, compare the key innovations intended to come out of VISION Cloud in order to ensure a good t of the research work and potential use. Media companies currently face challenges including more demanding consumers, novel distribution platforms, economic pressures, changing competitors (new content providers online, social networks, etc.) and different platforms of distribution. In order for established media companies to keep up with these changes and maintain their competitive advantages over emerging content providers, Cloud computing technologies should be considered. By enabling media companies to use cloud computing services either as cloud providers or as cloud users without compromising their content ownership and digital rights, many advantages can be obtained, including: faster time to market increased exposure to content (and possible increased sales) faster, fresher content packaged, identied, and available to the right consumer anywhere, anytime Furthermore, media companies can leverage cloud computing services 1 by: building digital archives in the cloud, applying best practices in storage and asset management developing strategy for standards-based metadata creation improving the efciency of content delivery optimizing costs of production processes supporting new ways of distribution

Among cloud computing solutions as broadly dened, storage clouds may have a specic impact on media companies in a number of different ways. For example they can:

be capable of efcient provisioning of storage capacity on-demand be able to tailor placement of data for efciency, locality, security and privacy restrictions. offer internal (to the company) or external (to the public) services breaking the barrier of storage
space limitations.

provide solutions for efcient processing of large data volumes in the cloud.
Structure and Approach of this market analysis The VISION Cloud project aims to contribute to cloud computing in the area of storage. This specic market analysis which is part of the media use case consequently asks: How can innovations in storage help in the Media editorial process from raw material to distribution? In order to get meaningful insights from this rst year market analysis, the conditions under which these advantages can be obtained must be examined from several points of view. The idea is to understand the needs of media organizations and connect those with functional as well as technical gaps of existing cloud solutions. Finding these gaps will help to tailor ideas and prototypes from VISION Cloud for future development and subsequent exploitation.

1 partially based on Hoehn, John A.: Cloud Computing for the Media and Entertainment Industry, June 2010, Somers, NY 3

White Paper: Media Clouds

The media organization point of view is structured based on a workow model dened by the VISION Cloud media use case partners. The goal of the model is to look at each step of the production workow, aiming to understand how certain use case scenarios can be applied In order to better understand the market analysis from a developers or technical perspective we added case scenarios currently discussed in White Papers and among experts such as Enterprise to Cloud, Enterprise to Cloud to End User, Enterprise to Cloud to Enterprise and Private Cloud. Examples of all of these use case scenarios can be found while going through the media workow. These models will additionally be described using more technical concepts such as Infrastructureas-a-Service (IaaS) or Platform-as-a-Service (PaaS). Summing up, in this analysis we want to understand the needs of editors and journalists on the one hand and then connect these with how developers would view a cloud platform that would serve those needs. In order to direct the work in VISION Cloud into areas where innovations are needed we examined best of breed solutions already on the market, which are used as benchmarks. This market analysis will be re-done and updated in the course of the project - both to understand how offerings have evolved and to further tailor potential outcomes from VISION Cloud.

White Paper: Media Clouds

2. Cloud Computing and Media Organizations


2.1. Introduction
An evolving context It is estimated that by 2012 there will be one trillion devices connected to the Internet and 50 billion of them will be mobile phones.2 This number is just one illustrative change creating growing pressure on media companies to change their processes and models. News organizations have to adapt to multiple changes occurring in the media industry. Author John Hoehn mentions: [T]he reality is that the consumer is in the eye of the storm 3 which means that while in the past content was scarce, now there is unlimited supply of content provided by traditional media companies as well as new media social network Internet sites via more and more portable consumer devices. Media companies are no longer gatekeepers of content or information. Consequently, media companies face, among others, the following changes: more demanding consumers novel and different distribution platforms economic pressures different competitors (new content providers online, social networks, telecommunication operators etc.) changing business models technical formats and protocols specic for various devices/platforms. In the same paper, Hoehn points out that Cloud computing represents many opportunities for media companies to improve their competitive advantage by getting content to multichannel, threescreen (or even four-screen) markets faster than ever before while potentially reducing costs.4 Media companies have to face new challenges in the consumer-centric environment. Media consumers do not care where the content comes from and, in fact, are getting used to receiving content whenever they want wherever they are on any device they use Needless to say, distribution latency plays an important role: If the users do not receive the content right away, they can be lost forever as customers, because theyll simply choose another media company as content provider.

2.2. External Challenges


Media companies worldwide are affected by the rise of digital content on the web. As a result there is a convergence process from older, analogue-based models of production and delivery to all digital, multi-platform strategy. Many commercial content producers have to cope with sinking revenues from advertising, sales or subscriptions in their core business. The switch to digital models is difcult, so far most media companies generate only a fraction of the revenues with their online offerings, when compared to the revenues generated in the past via traditional channels. As a result newspapers, book publishers, radio and TV stations are under pressure to compete
2 Hoehn, John A.: Cloud Computing for the Media and Entertainment Industry, June 2010, Somers, NY 3 Hoehn, John A.: Cloud Computing for the Media and Entertainment Industry, June 2010, Somers, NY 4 Hoehn, John A.: Cloud Computing for the Media and Entertainment Industry, June 2010, Somers, NY 5

White Paper: Media Clouds

with new entrants, ranging from global search companies like Google, start-ups and blogs. An additional development is a paradigm shift in the sources used by people consuming news. It is a complex, multi-level development. In general, a growing number of users get their news from the Internet, with increasingly the news being ltered through new social media platforms such as Twitter, Facebook or Google Plus. Information technology is expected to play a big role in how media companies can cope with this change, though the specic models and workows how this can be applied to digital newsrooms are not fully dened. There are early models, but the real challenge will be to get them working on a day to day basis. Additionally, there are many models to be considered; one size does not t all. A newspaper company will have a very different set of objectives than a broadcaster. At the same time, start-ups and social media platforms inuence how, when and where media content is consumed. Another driver of change is the shift to mobile devices. The question is: How will media companies manage to keep up with these changes and how do they secure their competitive advantages?

2.3. Internal Challenges


The digitization process has forced media organizations to adopt a wide range of specialized storage technologies to better manage the growing amount of digital media assets that are becoming more and more enriched (e.g. media + metadata). Internally many media organizations look for ways to transform their archives into a new tool that would not just be a library at the beginning and the end of the production process, but a platform that could assist in the time-driven production of content on a day to day basis. To provide just one example of the challenges: How could a digital archive enable effective use/reuse of digital media assets - making it easier to nd good shots, where the rights are cleared, etc. Cloud storage is one layer of this. The focus of VISION Cloud to examine innovative approaches to store any type of content in a cloud computing environment is what makes this particular project interesting for media companies. To clarify and to provide just one example, every phase of a media workow needs to re-examined here. For example, just take ingestion,i.e. when the audiovisual material generated by lming crews or coming from international feeds is imported into the media companys facilities. For the future this would require: 1)# 2)# # 3)# # 4)# storage with extreme high performance to allow multiple streams and a high throughput; storage with high capacity to meet the evolving needs of HD and HD stereoscopic productions; optimal placement of the content on the right type of storage according to how it is going to be used in subsequent phases; integration with a digital asset management platform to have the total views of assets;

As a direct consequence, content might be trapped in high performance storage silos. One open question here is how TCO (total costs of ownership) will actually map out. On the one hand there are clear benets and cost reductions in cloud storage. On the other hand media companies are facing considerable development and labour costs in this phase of transition to all digital workows. Ideally media clouds could provide two cost benets: Lowering the costs of storage and helping to reduce development costs including the time needed to search for certain content items in the journalistic workow. Collaborative access across multiple time zones, locations and organization requires content to be accessible anytime and anywhere. Therefore, a centralized viewing and simplied management of the total asset catalogue becomes crucial.

White Paper: Media Clouds

Additionally, there is more to media clouds than just storage at low costs. Other layers of such systems would include data protection and disaster recovery. As a result complexity increases as the size of storage system grows, further complicated by the proliferation of technologies which vary by storage system type. A new approach (partially covered in VISION Cloud as one innovation area) is the idea of computing in the cloud. In simple terms this could include the transformation of digital content for all platforms, from raw HD video material to versions more suited for mobile clients. Rendering and transformation for all platforms currently is one application in need of considerable computing power. These are not all, but some of the more obvious challgenges for media clouds. In the above outlined context, approaches for each of the media companies will vary. It is expected that cloud computing will prove a very helpful concept to achieve the objectives of delivering relevant content to pre-dened target audiences at a very fast pace. A cloud is an elastic execution environment of resources involving multiple stakeholders and providing a metered service at multiple granularities for a specied level of quality (of service).5 Also according to Hoehn6 , the following are a selection of cloud benets that have the potential to transform business strategy in many ways: providing globally available resources enabling real-time content and metadata sharing lowering total cost of ownership optimizing technology investments

To summarize, by enabling media companies to use cloud computing services in any way (either as cloud providers or as cloud users) without compromising their content ownership and digital rights, many advantages can be obtained. Among them are: faster time to market increased exposure to content (and possible increased sales) faster, fresher content packaged, identied, and available to the right consumer anywhere, anytime

2.4. Enhancing the digital supply chain with the media cloud
In order to deliver content efciently, media companies are already looking at the cloud to enhance their digital supply chain. Opportunities include reducing capital expenses in IT infrastructures and their management/maintenance and using new opportunities, such as economies of scale in content storage, media processing and distribution services. Demands for a media cloud are different from the demands for clouds in other industries since the media industry has unique workows, production systems and delivery objectives. A media cloud is a variation of a cloud, where all modules are optimized to address the needs of media companies. When enhancing their digital supply chain, media companies can leverage cloud computing services in a variety of ways7: building digital archives in the cloud, applying best practices in storage and asset management
5 European Expert Group Denition (in The Future of Cloud Computing. Opportunities for European Cloud Computing beyond 2010, January 2010), URL: http://cordis.europa.eu/fp7/ict/ssai/docs/cloud-report-nal.pdf 6 Hoehn, John A.: Cloud Computing for the Media and Entertainment Industry, June 2010, Somers, NY 7 partially based on Hoehn, John A.: Cloud Computing for the Media and Entertainment Industry, June 2010, Somers, NY 7

White Paper: Media Clouds

developing strategy for standards-based metadata creation improving the efciency of content delivery optimizing the cost of production processes supporting new ways of distribution

The one take-away here is that media and content are a distintive domain, where actual workows of journalism play key role. As a result media clouds might evolve as one distintive variation of cloud platforms in the future.

2.5. Evolving new models for media


One of the most insightful descriptions of what has changed comes from Jeff Jarvis8. The journalist/blogger/book author and Journalism professor has compared old vs. new models. One major difference is that it is much harder to dene the starting and end point of a production workow. Writing articles or producing videos and publishing them online is part of a dynamic process that can be enriched by user generated content and therefore can change over time9. To briey summarize the new way of producing news: Never starts, never ends Transparency of investigation Input & collaboration at all stages Enabled by links Enables networks This was easier in the more traditional models: Newspapers and TV stations used a distribution technology (printing presses, transmitters, cables) enabling them to reach a certain audience through a standard never-changing path. This created a value for advertisers, who could use the attention generated by news for advertising, and a much higher level of control and monitoring of business conjunctures and trends through the analysis of few well established indicators (e.g., audience scores). Today the picture is much more complicated: Distribution costs for web content have gone down to almost zero, content from a website can potentially be used instantly by users all around the world. The abundance of content offerings has created a situation, where media offerings around the world are competing with each other for the front row. As a result of these changes, this is what media companies expect from media clouds:

reliable storage transparent handling of multiple le types and formats, i.e. ability to automatically discover the key properties of content items without the need to upgrade the storage platform with third parties plugins; consistent metadata management across multiple copies, i.e. built-in capacity to cope with issues of data integrity, access security, xity of data items;

assistance with various processes of production, primarily by adding exibility and adding speed
e.g. through native provisioning of key content management and processing functionalities; usability low deployment efforts bandwidth optimization in distribution avoiding lock-in with any platform, e.g. ability to shift data to other providers, if needed

Furthermore, there are other requirements that media clouds need to fulll:
8

Jeff Jarvis, New Models for News (2009), Presentation via Slideshare http://www.slideshare.net/jeffjarvis/ new-business-models-for-news-presentation 9 Jarvis, Jeff: http://www.buzzmachine.com/2009/06/07/processjournalism/, 2009 8

White Paper: Media Clouds

Costs, handling and delivery quality: The main point is that there need to TCO assessments
comparing the total costs of ownership of media clouds to on-site datacenters. While this is not easy, providing more facts how media companies could lower costs and enhance the workows might help creating growing acceptance for such platforms. Agility: allows for fast changes, such as enabling access to an Application Programming Interface (API) to enrich content, ingest more metadata from formerly non-existent sources (e.g. Twitter, Flickr, Facebook) Competitive edge: As a result the convergence from traditional media workows to all-digital production one additional task is to provide a multi-platform distribution, to reach users on the web, tablets, mobile computers and other areas (e.g. cars, airplanes) Scalability: in case of breaking news or or broadcasting peaks in general, temporary bandwidth and storage requirements must be addressed In addition to what was mentioned above, one relatively new concept that has been added to reporting is data-journalism, where journalists use data as a source for reporting. This trend, while nascent so far, might have deeper implications if seen as an extension of business intelligence to public intelligence, where politics, companies and individuals would gain access to reporting based on deep analytics.

2.6. Cloud Architectures Relevant for Media Companies


According to a white paper produced by the Cloud Computing Use Case Discussion Group 10, there are several business scenarios that need to be taken into account. These models are helpful to translate the potential needs of media companies into more technical, standards-based views. The more common ways to describe these and other models of cloud computing are acronyms like PaaS (Platform as a Service), etc. Our focus in this section is how different concepts could be applied to media uses. We therefore look at ve main variations of cloud computing, through the lens of potential uses for media organizations. The ve main concepts are: End User to Cloud Enterprise to Cloud to End User Enterprise to Cloud Enterprise to Cloud to Enterprise Private Cloud The scenarios above share a certain technical infrastructure, but do all differ slightly from one another in requirements. In this section we want to briey explore why, how and when the models above are relevant for media organizations. 11 2.6.1. Enterprise to Cloud

Denition: A media company using cloud services for internal processes. Use Cases for media companies: using cloud for backup or storage of media content using cloud databases as part of an applications processing using various applications in the cloud (SaaS) (e.g. transcoding)

10 http://www.opencloudmanifesto.org/Cloud_Computing_Use_Cases_Whitepaper-2_0.pdf, 2009 11 the following screenshots were all taken from the Cloud Computing Use Case Discussion Group 9

White Paper: Media Clouds

2.6.2.

Enterprise to Cloud to End User

Denition: Media company uses a cloud platform to deliver data to the end user. Use Cases for media companies: provisioning of a datastore to the end-users (e.g. consumer services like remote PVR, calculators, consulting) distribution of content through content delivery networks for high-trafc/high-return publications

10

White Paper: Media Clouds

2.6.3.

Enterprise to Cloud to Enterprise

Denition: Two media companies using the same cloud in order to enhance the digital supply chain. Use Cases for media companies: managing copyrights and metadata across various enterprises; integrating the production process across several companies in the same business group or in different ones; integrating content during the distribution process (e.g., by using personalized advertising engines) 2.6.4. Private Cloud

Denition: The cloud is contained within the enterprise. Use Cases for media companies: CRM Payroll Production System

2.6.5.

Which cloud model for media companies?

This short overview shows the many options that are possible for uses in one or many layers of a media organization. At the same time, it must be noted that media companies are different from other industries - there is an emphasis on speed and timeliness on the one hand and a consistent need for changes in formats, process and output. Combining reliability of technical systems with agility when changes are needed is probably the biggest challenge here. In conclusion, media companies could potentially use many different models of cloud computing to their advantage. For example, a private cloud use case scenario is just as important as use case scenarios for enterprise to cloud, enterprise to cloud to end user and enterprise cloud to enterprise.

11

White Paper: Media Clouds

To add another perspective, which is again speaking in more technical terms: Both Software as a Service (SaaS) as well as IaaS and PaaS models offer opportunities in this space. Looking at potential benets of cloud computing variations, media companies should be aware that - based on the specic use of cloud computing - there are benets in mainly three concepts (or to use the acronyms: SaaS, IaaS and PaaS.) The specic benets that are possible are discussed in more detail below. SaaS can help media companies to minimize the upfront costs associated with hardware and software acquisition and helps companies avoid having to hire more staff to operate and maintain in-house systems. 12 IaaS can provide elastic or scalable infrastructure on demand to handle trafc and content overow situations. 13 PaaS can turn clouds into places to store, process, distribute and support exposure of content in an elastic fashion 14 What is needed are cloud based infrastructures which have already adapted to typical media uses - in order to speed up the process of adapting such systems to a specic company.

12 Hoehn 13 Hoehn 14 Hoehn 12

White Paper: Media Clouds

3. Role of VISION Cloud


3.1. About VISION Cloud
VISION Cloud, a three-year project partially funded by the European Commission's Framework Programme 7, is researching ways to build a scalable, exible infrastructure for efcient handling of the large amounts of data stored in the cloud. The VISION Cloud architecture aims at providing a reference model for a storage cloud which will have low provisioning costs, be highly efcient in terms of bandwidth requirements, have signicantly lower maintenance costs, and overcome many of the issues preventing many businesses from transitioning to the Cloud, namely in the areas of security, reliability, availability and cost. In particular, VISION Cloud has identied some main areas of innovation: raising the abstraction level of storage to support easier and more efcient access to information data mobility and federation across clouds to overcome vendor data lock-in computational storage, which brings computation close to the stored data content-centric access to stored data which will make content visible to users, as opposed to the underlying storage container mechanisms for cost-efcient storage with Quality of Services (QoS) and security guarantees. VISION Cloud was conceived to address specic requirements coming from a wide range of use cases, including media, telecommunications, enterprise applications and healthcare.

VISION Cloud has two participants in the media use case:


RAI (Radiotelevisione Italiana), known until 1954 as Radio Audizioni Italiane, is the Italian state owned public service broadcaster. Deutsche Welle is Germanys international broadcaster: online, on-screen and over the air. It provides a European perspective to audiences around the world and promotes intercultural dialogue. Several innovations from the VISION Cloud project can help meet the requirements for a media cloud15. However, the following in particular, address the media specic requirements: Content-centric storage with its potential ability to allow accessing content items (e.g. audio and huge amounts of data) stored within VISION Cloud through the specication of metadata and relations among items. Computational storage, and its potential ability to move processing of the content (e.g. audio and huge amounts of data) near to the data, thus relieving services from possibly heavy computations, and, as the main part of the computation will be done inside VISION Cloud, reduce the network bandwidth needed. Through the full development of the above mentioned features, VISION Cloud will thus: offer lightweight services running on a large variety of devices and offering the same capabilities regardless of the access means; ease the development of new services, reducing time to market and allow third-party services to be modularly integrated. offer services breaking the barrier of storage space limitations. provide solutions for an efcient processing of large data volumes in the cloud be capable of efcient provisioning be able to tailor placement of data; for efciency, local, security and privacy restrictions.

15

based on Nessi et.al.: Exploting Cloud Computing - A Proposed Methodology for Generating New Business 13

White Paper: Media Clouds

4. Market Analysis: What are the best solutions so far?


As the VISION Cloud project continues to evolve, providing a reference model for a storage cloud, this market analysis will be in need of being updated over time. But as of late 2011: Which are the best solutions for cloud-based media workows? The idea here is to identify the platforms that work for a certain phase. Our motivation was to rst look what is already there, understand why the particular solution is helpful and then analyze how the best features of what is there today could potentially be applied to cloud storage, which is the main focus of this particular research project. The approach aims to describe why certain platforms and cloud computing offerings have already gained market traction and why they are usable for media organizations, based on the requirements described in preceding chapters.

4.1. Basic model: A cloud-based media workow


In order to structure this market analysis, we will use a cloud-based media workow dened in the VISION Clouds media use case. The goal is to look at each step of the production workow, aiming to understand how certain use case scenarios can be applied. The focus here is on how to handle large amounts of content in each phase, from initial production to publication. The workow for media use cases in VISION Cloud includes eight distinctive modules or steps: Team, Ingestion, Filter/Metadata, Production, Publishing, Distribution, and Popularity (see table below for more details). The one thing that is important here: Media companies will have an increasing need to handle ever larger amounts of content. The rise of HD video, the need to understand big amounts of data and several other factors will increase the number of les, the number of versions of one le and as a result will affect the workow in every newsroom. Media companies are driven by time, there are daily deadlines to meet in most organizations. Therefore, the emphasis here is to translate key concept such as Software as a Service and varionts of cloud computing into applications and services that are supporting the work of journalists. This is what this workow model can illustrate. For each phase we ask: What would be a compelling and innovative solution, where cloud computing could create a new ow? Workow Vision Cloud (Media Use Case, 2010)

To illustrate, the phases of this workow can differ greatly in terms of effort, time and costs needed for each one. The following visual aims to show how - using the basic model above - it would be possible to map the process to different media companies, depending on which area is the most important and even taking into account different strategies of print, broadcasting or pure online offerings. Flexibility of the Vision Cloud media use case (example showing how the role of each phase might vary)
14

White Paper: Media Clouds

Each phase of the media workow is a chain of different tasks that need to be taken into consideration and require different delivery models and include numerous cloud computing use case scenarios, as illustrated in Table 1.
Table 1: Media workow vs. cloud use cases scenarios and delivery models.

Phase

Main task

Collaboration/ teamwork 1. Ingestion

2. Filter

3. Production

4. Publication 5. Distribution

Cloud Computing Use Case Scenarios Setting up users, Many media projects involve different N/A basic collaboration, people at different phases. status. Uploading material Wide variety of le types and formats Enterprise to into storage. are typical for media (HD video, Cloud, photos, images, audio, code). Enterprise to Cloud to Enterprise Describes a stage A good lter function is key for search. Enterprise to where material is This is usually not possible by Cloud, Private structured, labeled hierarchical storage. Cloud, Enterprise and enriched with to Cloud to metadata. Enterprise Editing the material, Flexibility needed and ability to add Enterprise to often software (e.g. APIs to send/receive material which is Cloud, Private Avid for video, then stored in the media cloud. Cloud visualization tools for data). Providing access Better data and medata strucute of Enterprise to for the public. published pages Cloud to End User Enabling Multiplatform devices (web, TV, Enterprise to publication to tablets, smartphones, phones). Cloud to End User multiple platforms.

Notes

6. Popularity/ Measuring reach and success

Usage data, such as visits, views, comments, embeds, tweets, etc.

Need to see both quantity and quality Enterprise to (e.g. inuence, engagement). Cloud

15

White Paper: Media Clouds

4.2. Workow: Teams


Workow Step 1: While Teams cannot really be seen as one phase of the production cycle, one has to consider effective collaboration before every project. Media content is produced by many people, the better a system would support each task, the easier the workow might e. While setting up users for a project, which is provided by most software packages, few offerings so far have ways to provide a good mapping of the process from a single user view and a team view. An example or a test case would be a digital archive, where certain team members work in different steps of the workow. Better collaboration of teams, supported by innovative cloud platforms and storage, might be on important aspect of media clouds. Today, many higher level workows (e.g. for videos, TV or data journalism) are executed on one PC, then the material is handed over via mail or services like Dropbox oder delivery services like YouSendIt (as many mail systems to not accept les sized above 10 MB). This is time consuming and error-prone. Re-use of material is often limited, a lot of potential to work faster can not be realized. Benchmark: Dropbox Dropbox, a seemingly simple offering allowing users to create a shareable, synchronized folder to exchange les, has become extremely popular in the last two years. Driven by its popularity and wide user base the company is now valued between two and ve billion US-Dollars, after receiving only about 7,2 Million US-Dollars in the start-up phase.16 Although there are many competing services, the simplicity of the service has made it a winner over several older offerings, e.g. YouSendIt. An overview of competitors can be found in the CrunchBase prole of DropBox.17

4.3. Ingestion
Workow Step 2: The second phase in the media use case it Ingestion. It is the process of uploading any material into a storage/production system used by a media company. Good ingestion of les and metadata is crucial for the quality of any archive. Currently, most storage systems or production systems have a limited handling of metadata. They either have a schema of their own or they lack connectors to ingest available metadata, such as EXIF-data from cameras or usage information from big platforms like Flickr. For example, when importing a photo from Flickr, the original source, the owner and the rights status are easily lost. This in turn might create considerable extra work or doubt whether an item can be re-used or not. Scaled to thousands and millions of content items, low quality or even inconsistent metadata creates a problem. Metadata is central to the management and organization of resources using ResourceSpace. Many other DAM systems place resources on a tree type structure, as with a typical le system on a PC. This works to a point, but it only allows each resource to be classied on one way. Specic requirements for data and content ingestion:

16 17

http://techcrunch.com/2011/07/12/dropbox-raising-massive-round-at-a-5b-plus-valuation/ http://www.crunchbase.com/company/dropbox 16

White Paper: Media Clouds

Ease of use: while rening the data is part of the next step within the workow outlined in this
paper, the data needs to be easily transferred from the source into the cloud. Part of a qualitative ingestion is the usability part: many reporters, journalists, editors or archivists might want to ingest new data - passive and active data - and the user experience is of importance. High Perfomance Storage: When talking about real-time data, live coverage of events is relevant. Therefore, the speed of being able to ingest the data and working with it is crucial. Rich Data/Metadata: Needs to be able to work with all kinds of metadata for active as well as passive data. Scalabilty: to meet the evolving needs of HD and HD stereo production, avoiding traditionally under-utilised and over-provisioned storage strategy. Smart storing: optimization of the content placement on the right type of storage according to the rules it is going to be used Compelling data protection integrity and security and disaster recovery strategies

Benchmark: ResourceSpace (http://www.resourcespace.org) This Open Source tool is free to use, download and can adapt in many ways to handling ingestion. ResourceSpace is a Digital Assets Management tool, which is a sub-genre of Content Management Systems. The reason why we favoured this tool over other, commercial versions is that is has a proven record to handle les with rich metadata. And, while not developed explicitly for cloud usage, it might have the right size plus the adaptability for most newsrooms. Bigger systems, developed for video and broadcast needs might have some extra benets, but as such ResourceSpace offers a lot of freedom to experiment and still start based on highly developed system with a proven track record. The software enables teams to organize content, search for items and thus provides a hub to upload, download and further process digital les. Based on its original focus in advertising agencies or companies in need to handle multiple media assets, it provides the needed reliability for different projects with different deadlines. With ResourceSpace, the use of metadata as the principal method of storing and nding the content is highly qualitative. In addition to that, the option to create content sets, that can be shared with others to work on a particular project, is very useful. The software even supports versioning, meaning that there is a way to distinguish older from newer les (which is a real challenge and must-have feature for audio interviews, videos and datasets). The original purpose of a DAM is to help an agency or other media companies (here: mostly advertising) to collect and use all material needed for big campaigns, such as photos, videos or advertising displays. Such collections can become quite vast very quickly, e.g. 20.000 photos and numerous variations within a short amount of time. This software can provide an excellent middleware or interface for cloud storage. Being written in PHP, with good documentation, there are many options to adapt this software for media workows and raw material ranging from datasets to photos to video. Given the wide availability of PHP developers much of that work can be done in-house. Installed on a platform with sufcient storage capacity it should perform even when loaded with very large amounts of les.

4.4. Filter, transform & search Workow Step 3: Once material is ingested, there is a need for an in-between step currently
creating much manual labor. Content must be ltered, formats must be transformed to create a solid base for reliable search. There are three main aspects:

Standardization: standardized ways to structure data in order to skillfully visualize and offer
them to the audience in a certain way
17

White Paper: Media Clouds

Compliance: EU Metadata standards 18 (e.g. EBU Core19) have to be taken into consideration Rich Data/Metadata: Ability to annotate the data in order to complete and enrich stored les
Benchmark: Google Rene 20 Google Rene enables users to clean up big sets of data. This powertool has not been developed for media or cloud computing, but provides a new approach to clean up data, especially for storage in clouds. Google Renes excels at identifying terms that might mean the same, but are written differently. Google Rene is not capable of automated handling of such processes, but provides a recording function to run the same processes on datasets as they come in. As of today though it could be used as a method by journalists, archivists to create a solid and clean base of metadata. It allows journalists or anybody in the media business to work with messy data by importing data and opening it up in a spreadsheet. Functionalities include ltering and clustering. Furthermore, it allows users to share documents which means it allows more than one user to work with the data together. Also, one can merge more than one data set together.

4.5. Production

Workow Step 4:

Production in the context of a media cloud is the transformation of raw material into content. In digital workows a wide variety of input formats can be used: Data, text, pictures, raw video material, links to other sources, etc. The production process can be part of a media cloud (e.g. a connected Content Management System) or through usage of external tools such as Final Cut or other software for video editing. The main challenge of this process is to balance quality and quantity, for publication of multiple formats for multiple channels. Requirements:

Ease of use Digital Asset Management system: it is of utter importance that usability, conguration and management of interfaces are taken into consideration as well as an easy way to import, share and export data across the organization Computational Storage features: the production process is the most complex phase that may require high and fast performing intensive operations (e.g, rendering, transcoding,..). Time is a crucial factor, especially for media companies who produce news Scalability: Within the production phase, journalists, archivists, editors and engineers have to be able to work with different types of formats as well as with different sizes of formats. The cloud-based services need to be able to work with low-bandwidth material as well as with HD and high-bandwidth content. Collaboration and reuse: accessing digital media assets across multiple time zones,

locations and organization, anytime and anywhere, with minimal time and effort
Benchmark: Edit/share Edit/Share21 is an open-source, online video editing software. It includes options to work with AVID and Final Cut, two of the leading video editing software suites. Features of Edit/Share:
18 19

http://en.wikipedia.org/wiki/Metadata_standards http://tech.ebu.ch/lang/en/MetadataEbuCore

20 http://code.google.com/p/google-rene/, further important platforms include DataWrangler, Open Calais, Box.net and DocumentCloud
21

http://www.editshare.com/index.php?Itemid=208&option=com_wrapper 18

White Paper: Media Clouds

platform that provides ingesting, managing and storing options for multichannel broadcasters product includes Geevs Video Servers for ingestion and playout product includes a Flow database that allows users to track and manage all media assets throughout the production process product includes a scalable storage

Additional features: allows users to connect any number of storage capacities of over 200TB allows AVID and Final Cut Pro users to with same les (MXF, etc.) if necessary backs up every clip as soon as done capturing allows editing while capturing includes administration features for big organizations bandwidth-controlled le ingest end-to-end production and playout workow

Other options that can be used during the production phase. 4.6. Publishing Workow Step 5: Publishing on digital platforms aims to make available all forms of content.
One new trend is that the boundaries of media are vanishing, e.g. broadcasters increasingly publishing texts as well as newspapers publishing large amounts of videos on their sites. Moreover, publishing has started to include not only webpages but also applications due to the widespread diffusion of smart phones/tablets and connected TV. The tools to both congure and publish content are Content Management Systems (CMS) which come in wide variety ranging from free to commercial solutions used by large companies around the world. Need for new publishing models and technologies The current model of publishing webpages, which are effectively a mash-up of code, links, text and many other media content types is under criticism, there is a quest to nd new, better models both in terms of usage as well as from a technology point of view. Requirements: Format (resolution, encoding, wrapping, ) conversion on-demand Adequate metadata formatting according to users device Advanced ltering and searching capabilities Versatile and exible data model Benchmark: YouTube We have chosen YouTube as a benchmark for cloud based publishing mainly because it shows how extremely high amounts of les can be uploaded, stored and delivered to a global user base. While there are many features lacking in terms of enabling to structure and control content on the platform, the main point here is that YouTube can serve as a model in terms of scalability and robustness for global publishing of content. Arguably the back-end of YouTube is the fully scaled CMS based on cloud storage, exceeding the needs of even the biggest media company that there is in terms of storage and bandwidth.

4.7. Distribution
Workow Step 6: Distribution is the process to deliver content to multiple digital platforms, potentially including feeds from or to print media, broadcast media. In fact, distribution is changing from a one-way delivery process (e.g. newspapers being distributed to subscribers) into a versatile hub for digital content.
19

White Paper: Media Clouds

Using APIs distribution can serve both for sending out or ingesting content (including feedbacks and multimedia contents provided by nal users). A newer addition to distribution is to include notications or sending links to social media platforms, either by manual work or automated processes. Moreover adaptive delivery of multimedia content according to the available bandwidth and terminal capabilities has to be taken into account. Benchmark: Signiant 22 In our view this US company has developed a whole suite to address technology needs for reliable distribution. The company offers a control center, where workows and format transformations can be dened. Furthermore the company offers ways to optimize delivery of video content over available bandwidths. Notable other alternatives Managing digital distribution has created a whole industry with different players. This would include both hosting companies as well as content delivery networks powering big websites around the world, e.g. Akamai. Offerings like the suite from Signiant above though would connect to these other layers of the infrastructure at dened points. API Platforms and Mashups Application Programming Interfaces have already been mentioned in the ingestion phase. But the process of automated, rapid exchange of media les and data works both ways. So, APIs can be used as an increasingly important aspect of distribution, too. Here, again, the quality of the data, especially the metadata is crucial. Creatively used, APIs can be combined and enriched with other data (Geotags, etc.) and then used for new aggregated offerings, for the creation of maps, and so on. APIs: Cloudmade, Programmable Web, Mashery In this emerging model a media brand is not building its reach just by having a website. Instead, by making branded content available, users can nd content on other platforms, subscribe to it, comment on it, and embed it. These multiple uses are changing traditional models of media profoundly. There is no xed time when to use the content - be it photos, audio podcasts, videos or data. Furthermore, there are new ways to combine content form several sources, ltering content for certain interests, etc. Next generation aggregators: Flipboard Based on API access, new business models can be created - ltering news to a users liking, based on topics of interest or location.

4.8. Popularity aka Usage Analytics Workow Step 7: Measuring the impact of content has always been challenging. Before pointing
out the challenges of todays digital media in terms of accuracy, it should be remembered, that the whole advertising business has been working based a very limited ow of information. Print Magazines and TV invented a wide array of how to describe the reach and distribution of a media product, often relying on small panels of users or surveys. Benchmark: Google Analytics While the platform has issues in terms of privacy and potential undetected use of aggregated data, the free offering has many features, can be integrated into other platforms and is robust. It must be noted, that there are by now alternatives to Google Analytics, that are Open Source and handle data differently (e.g. do not store IP-based data of users on servers in the US, which is an area of jurisdictional debate at this time in several European Countries).

22

other notable platforms include SaxoTech, CCI, Atex and Woodwing 20

White Paper: Media Clouds

Google Analytics is a powerful tool to track usage on web pages. Its free, easy to deploy and can be used to track a number of different projects over time. There is no learning curve, even the implementation of the tracking code is very simple. However, there are also a few drawbacks. One thing is that Google Analytics is free for a reason. In default mode the tool will pass on data to Google, where the incoming usage statistics from millions are used to rene the algorithms for advertising. While this is not negative per se, there are some practices that are violating data security rules of European countries. Every site using Google Analytics should have a data policy in place, many have not. There is a option to not sent any data to Google. But as Google is very dominant in search and given the fact that users can be 100% sure of where data is being recorded, it is advisable to look for alternatives.

21

White Paper: Media Clouds

5. Early conclusions: Media Clouds - what is there and what is missing?


The VISION Cloud project aims to contribute to cloud computing storage concepts. From a media point of view the most interesting area of work is content-centric storage, meaning that les stored in a platform can be worked with in innovative ways. The key challenge to achieve this is a better, more reliable concept for metadata that is attached to the les in combination with the specic drivers inside a media organization. There is a need to nd ways for better and more extensive ingestion of les into storage as well as nding innovative, reliable and potentially automated concepts to enrich media les continuously. One video le could be updated throughout its use cycle with basic metadata (such as: Date of recording, resolution), with usage data from the organization (rights, last use, number of times this le has been used, etc.) as well as popularity data (number of views, number of Tweets, number of embeds and so on) In this view it can be debated whether we are talking about a new production workow - or alternatively - about ideas to create an active archive, meaning that the archive is not an endpoint, but moving higher in the workow become a hub for digital media production. Probably a new wording must be invented to distinguish this digital workbench from what we so far view as an archive. A more segmented view could be an architecture for cloud storage where the archive is the base layer, interacting with all stages of the production process, e.g. a production system providing dened access to stored les or equally a distribution system which interacts with storage when e.g. a re-attribution of rights for millions of les is needed. Looking back, numerous projects undertaken to create better, higher quality and more consistent metadata have not created a breakthrough so far. One problem is that if the metadata quality as a whole falls under a certain threshold the systems are deemed not reliable. Often this results in low usage. We hope this White Paper provided some new views on why Cloud Computing creates a eld of innovation opportunities. These are our early, rst year views on what concepts are debated and which solutions are already out there. If you would like to comment, get in contact or discuss alternative ideas to create better media workows, visit us at: http://www.visioncloud.eu or contact us via mail (information provided on the website).

22