| icnrg | L. Wang |
| Internet-Draft | L. Geng |
| Intended status: Informational | China Mobile |
| Expires: January 3, 2019 | July 02, 2018 |
Consideration on Applying ICN to Edge Computing
draft-wang-icnrg-icn-edge-01
Aiming at research on applying Information Centric Networking (ICN) technology to Edge Computing, this document analyzes the reasons and opportunities of applying ICN to EC. As well, towards this end, technical considerations are described and relevant scenarios are shown in the document. Benefits of deploying ICN at edge is analyzed in the document.
This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.
Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.
Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."
This Internet-Draft will expire on January 3, 2019.
Copyright (c) 2018 IETF Trust and the persons identified as the document authors. All rights reserved.
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.
Information Centric Networking (ICN) takes significant technical revolution and fundamental change on communication and networking. It uses content/information centric networking to replace traditional address-centric networking which change the existing networking model essentially. It can also be regarded an Internet structure evolution from host-centric structure to data-centric structure which means accessing data by naming. This structure enables to make the data relating application more independent of its location and transmission method. What’s more, security mechanism is based on information instead of host and the caching in forwarding process that promotes huge information transmission efficiently. It is very promising to apply ICN to some popular network architecture.
Meanwhile, Edge Computing (EC) is becoming important network architecture because of its outstanding performance in real-time, reliability, security, etc. It deploys services on the edge of network to be close to consumers, and offers decentralized function to enable excellent properties in local computing, storage, connectivity and so on. At present, Edge Computing works broadly on IoT and industrial verticals such as Energy, Manufacturing, Smart City and Smart Grid.
Therefore, it is worth attempting the possibility of using ICN on EC. ICN naturally supports decentralized caching, self authentication and multicast that can enable EC deployment. The combination of ICN and EC is able to offer a win-win approach and benefit mutually for maximum performance. In the following sections, we will seek the opportunities of applying ICN to EC, and outline the correlative properties of both. The technical consideration of the approach and relevant scenarios will be described as well.
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in [RFC2119].
EC- Edge Computing, an network architecture that provides local compute, storage and connectivity services
Other ICN related words used in this document are interpreted as description in [ICNRG-Terminology].
In traditional networks most typical service nodes are deployed in centre, so the network flow are transferred from the centre to the edge and downlink traffic is dominant. But as the IoT highly developed, a large amount of devices are deployed on the edge, which, therefore results in considerable uplink traffic. The requirement of traffic service flattening requests the technique that is able to make local communication for traffic convergence. This could be the entry point of applying ICN to EC.
Both ICN and EC possess some correlative properties, such as decentralized deployment, local communication capacity, producing abundant uplink traffic flow, etc. However, there are also some other properties they posses respectively which are complementary.
For ICN, caching and forwarding are two basic functions which are more about connectivity. But in practical cases, ICN node devices such as gateways demand for light computing and storage functions as well. Light computing and storage can make the network more dynamic, flexible and enable some AI deployments as well. Fortunately, edge computing is able to support storage and computing naturally. A combination of both ICN and edge computing can be mutually benefitted for maximum performance.
TCP/IP network model has been used for quite a while and is worldwide deployed now. No matter according to cost, difficulty, risk or other consideration, it is not realistic to deploy ICN on the whole network. However, the partial deployment of ICN can have a chance, such as ICN over IP or IP over ICN. Deploying ICN on edge service not only can help to mitigate the ICN whole-network deployment complexity, but also makes the network model more flexible.
In some scenarios that the network is not able to offer end to end communication such as power failure or natural disasters that could result in the interruption of the network and other local disconnections problems. Often such failure can cause a series of accidents and even chain reaction, resulting in the loss of enterprises and production. In the case, edge computing enable to supply service which is closer to the edge of data generation and business control deployment, making the computation much closer to the data source. Even if the network fails to get connected, the device can rely on local networks for data communication and processing.
However, (a) sometimes the data stored on the edge is “staleness” due to incapacity of updating timely. This could result in the mistake or staleness data transmission. (b) Furthermore, the storage space on edge is limited. For instance, it is not able to update new content if there is no spare space when storage on edge which normally is small storage capacity, is full. This can also cause the (a) problem.
In ICN network, the content is cached along the path it delivery. So the objective content can be from the source node or the other content caching nodes. When the network is disconnected, the caching content or data in decentralized nodes can be used in edge computing. Caching algorithm of ICN is able to solve two problems stated in previous paragraph by updating data efficiently and dynamically. This benefits from the caching replacement policies of ICN. The policies, such as LRU or LFU, provide mechanism how long or how often the data will be updated. Therefore the data is either the newest or the most popular. Decentralized content caching of ICN strengthens EC network disconnection solution and make more flexible networking.
In IoT industry, there are a huge number of devices deployed on the edge which result in a significant amount of uplink flow traffic. In EC, the prominent quantity of traffic is easy to cause traffic congestion.
In ICN network, the data content not only from the source node, but also it is cached in other nodes along the delivery path. So when the edge node request the data, it is not necessary to deliver data from the source node. For instance, in the figure, if node 1 is source node. When node3 requests data from node1, the content will be cached in both node A and node B. So next time when node4 needs the same content data, node B will deliver it, and vice versa. In the case, the traffic is not from node1(source node) to node3 or node4 anymore, but mainly from A to B. Therefore, ICN decentralized content caching enable traffic convergence.to reduce traffic congestion.
Data Traffic
+------------------------+
| |
+----+ +----+
+-----+ A +-----+ +----+ B +------+
| +----+ | | +----+ |
| | | |
| | | |
| | | |
+--------+ +-------+ +----+ +------+
| node 1 | | node 2| |node3 node4 |
+--------+ +-------+ +----+ +------+
Source Node
Figure 1
Traffic Convergence in ICN
Security problem is crucial and urgent to the EC applications. Firstly, there are many devices on edge are exposed to users which is easy to be attacked. Secondly, although authority level on edge is lower than host and cloud, there are more people can get access to the devices and application. This is in consideration of the consumer convenience and deployment flexibility. Hence, application and services are vulnerable on edge.
Instead of binding security to host node, ICN advocates the model of trust in content. This offers host-independent security mechanism which focuses more on securing information object and content trust. It means host attack no more can interfere edge application. Furthermore, self-certify names model of ICN enable to verify the binding between public key and self-certify name in distributed system without relying on a third party. This can reduce the security risk of involving a third party.
No matter ICN or CCN, they all promote content centric communication model. Independent from host node, naming on edge node gain more valuation on edge devices.
In consideration of cost and complexity of deploying ICN, it is not necessary to use ICN in the whole network. ICN using on edge is enough to highlight its advantage. Furthermore, there can be a corporation between ICN edge service and IP network.
Content Delivery Network (CDN) system, based on IP, composes a couple of servers that deliver content to a user, based on the geographic locations of the user, the origin resource and the CND server nodes. Normally, the resource is distributed in a downlink traffice in figure2.
However, in a ICN network, resource or origin node is not the central node anymore. An edge device can be the origin node that provides the resource which is delivered to ICN servers, and further distributed to the receiver nodes. As a consequence, the routing is from edge to central, or there will be an uplink traffic. This just reverses the CDN Mechanism, which is shown in figure3.
Therefore, deploying ICN network at edge that pulls the requested data from resource edge node to the servers can cooperate with CDN network. An ICN/CDN server is anticipated to translate the protocol between both and deliver the data to the receiver nodes which can be ICN network or IP network.
+----------+
|Origin/Web|
|SerVer |
+---+------+
|
+-------------------+ Traffic
| | | from origin to edge
| | | downlink
| | |
| | |
v v v
+---------+ +-------+ +------+
| CDN ++ | CND | | CND |...
| SERVER| | SERVER| |SERVER|
++---+----+ +--+----+ +---+--+
| | | |
| | | |
| | | |
v v v v
+-----++ ++----+ +-+---+ +---+---+
|EDGE | |EDGE | |EDGE |...| EDGE |
+------+ +-----+ +-----+ +-------+
Figure2 CDN Mechanism
ICN/CDN SERVER deliver data to receivers
+----------+ +------+
+---------->+ ICN/CDN SERVE+---------->+Receiver
^ | | | +Node--+
| +----------+ |
| | +------+
| +---------->+Receiver
+----+-----+ +Node--+
| ICN server
+-+--------+
^ ^
| | Traffic from Edge to ICN server
| | uplink, reversed CND
| |
+---++ +-+--+
EdgeNode|Origin | |
|Resource| |
+----+ +----+
Figure3 ICN Mechanism
This draft described the correlative properties of ICN and EC to analyze the opportunity of applying ICN to edge computing. The traffic uplink flow model is the entry point of this research. We could see ICN deployment is beneficial to EC by combining the outstanding performances of both. Furthermore, a win-win model is schemed in the document by means of mutual complementing. However, there are still challenges on deploying ICN on edge such as high speed mobility, fast context resolution and so on. These questions need to be answered in the future.
| [I-D.boucadair-connectivity-provisioning-protocol] | Boucadair, M., Jacquenet, C., Zhang, D. and P. Georgatsos, "Connectivity Provisioning Negotiation Protocol (CPNP)", Internet-Draft draft-boucadair-connectivity-provisioning-protocol-15, December 2017. |
| [ICNRG-Terminology] | "Information-Centric Networking (ICN): CCN and NDN Terminology" |
| [RFC2119] | Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, DOI 10.17487/RFC2119, March 1997. |
| [RFC5440] | Vasseur, JP. and JL. Le Roux, "Path Computation Element (PCE) Communication Protocol (PCEP)", RFC 5440, DOI 10.17487/RFC5440, March 2009. |