As the volume of content (data, services, applications, etc…) available via the Internet and other sources continues to expand, there is a growing interest in creating a network architecture that can scale network traffic at the rate of content creation, rather than the rate of content consumption – as is the case in current network architectures.
To fully understand and appreciate just how much demand there is for content and applications, look to a few data points such as Netflix in which its subscriber base of more than 93 million watched over 125 million hours of content per day in 2016[1]; or Apple, where the number of apps available in its App Store exceeded 2.2 million as of January 2017[2].
From a traffic perspective, the volume of video alone is expected to grow from 68% of global consumer Internet traffic in 2015 to 82% by 2020. Cisco’s latest Global Cloud Index Report[3] further shows annual global data center IP traffic growing from 4.7 zettabytes per year in 2015 to 15.3 zettabytes in 2020. Compute and collaboration, streaming video, social networking and IoT are some of the applications driving this growth.
For now, the current network architectures are supporting the volume of content; however, this is not without complications, inefficiencies and security challenges.
It is important to remember that the network was designed primarily as a communications network. A number of solutions have been implemented over the years to evolve this network such as content delivery networks, load balancers, open caching, moving content closer to the edge and transparent proxies, as well as a number of TCP/IP enhancements. However, the network remains connection oriented, requiring a connection from the server to the client for each data transfer.
The inefficiencies of the current network include its point to point nature and its limitations to utilize broadcast; communication that requires both parties to have unique IP addresses; as well as the fact that the user is unable to utilize local machines if a global connection is unavailable.
Finally, from a security standpoint, the current network places trust on the host with which the client is communicating rather than on the content itself. There is an opportunity to improve overall security by validating the content — for example, by treating the content as a data object, not tied to any particular host, and assuring the content is signed by a publisher.
How do we meet this content growth challenge?
The following market and technology trends are helping to shape the network’s future state:
To address the new demands that these trends will place on the network, ATIS recently launched its Evolution to Content Optimized Networks (eCON) Initiative. eCON offers guidance and actionable steps that will help establish a path from today’s IP-based routing network to a future network that leverages the increasingly important role of content.
Under the future eCON architecture, the network evolves from a host-centric dependency to one that is based on treating content as named data objects that are no longer tied to a particular host or address, but may be present in multiple locations across the network. A key differentiation between current networks and what is often referred to as Information Centric Networks (ICNs), one of the primary architectures assessed by eCON, is that the content object is delivered to the consumer by name, without consideration of where the requested object may be found at the time of the request.
For example, a consumer asks for content by broadcasting their Interest over all available interfaces. Any node hearing the Interest and having data that satisfies it can respond with a data packet (content chunk). Data is transmitted only in response to an Interest packet request, and a “content store” fulfills the request.
Because both Interest and data packets identify the content chunks being exchanged by name, multiple nodes interested in the same content can share transmissions over a broadcast medium using standard multicast suppression techniques.
From a network efficiency standpoint, an Interest packet is only transported through the network as far as necessary to find the named information. This could be one network hop away or several.
Opportunities & use cases
ICN is a networking architecture built on IP’s engineering principles, but using named content rather than host identifiers as its central abstraction. This approach retains the simplicity and scalability of IP while also offering better security and efficiency. Although an ICN architecture can be viewed as a disruptive evolutionary path from IP, it will most likely be deployed in an incremental fashion as an overlay, allowing its key advantages to be applied on an application or use case basis.
ATIS has identified potential target opportunities and an initial set of use cases that could leverage an ICN architecture or some elements thereof. These include:
- Linear TV/IPTV
- Efficient Automated Edge Caching
- Seamless “Session” Management and Continuity
- Integral Multipath Support
- Internet Broadcast/Multicast Services
- Object Based Communication Models for IoT Devices
- ICN-based A/V Conferencing
Summary
Presently, there are a number of research and development activities around the concept of named content networking solutions and the ICN architecture in both academia, government and industry-based technical organizations such as IRTF, IETF, ITU-T SG13 Focus Group IMT-2020, 3GPP, and NIST to name a few.
Although much of the industry standards work on named content solutions remains in a developmental stage, the ATIS eCON Initiative offers an initial assessment of early deployment opportunities and uses cases that could benefit network providers and the broader content industry.
Key advantages of this future architecture include: better security – where the data itself has been secured; more flexibility, allowing data to be location-independent and more adaptable; better network efficiency and performance; and a more scalable network that can better address the growing amount of content by enabling caching, native multicast traffic, native load balancing and more effective resource planning.
Learn about this new technology in greater depth in the ATIS white paper, Evolution to Content Optimized Networks.
Sources:
[1] Source: Netflix, Overview. Accessed February 22, 2017 from https://ir.netflix.com/
[2] Source: Apple, Q1 2017 Results – Earnings Call Transcript. Accessed February 22, 2017 from https://seekingalpha.com/article/4041266-apple-aapl-q1-2017-results-earnings-call-transcript/
[3] Source: Cisco, Global Cloud Index Report. https://investor.cisco.com/investor-relations/news-and-events/news/news-details/2016/Cisco-Global-Cloud-Index-coProjects-Cloud-Traffic-to-Nearly-Quadruple-Representing-92-Percent-of-Total-Data-Center-Traffic-by-2020/default.aspx
