User:Nakon/Identifier Network

The Identifier Network   was first proposed by Professor Hongke Zhang, a chief scientist in China's National Basic Research Program (Program 973).





Overview
Identifier Network is novel design for the future Internet architectire (FIA). Different from the traditional 7 layers-based OSI (Open System Interconnection) network model, the Identifier Network includes two layers：the infrastructure layer and the pervasive service layer. The infrastructure layer is used to complete all kinds of network terminal access, forwarding and switching routing of data information in network, providing unified communication platform for network services such as data, voice, video and so on. The pervasive service layeris used to complete unified management of service data resource and unified control of network connections, so as to structure pervasive network service architecture and meet the requirement of different network.

The "four identifiers" include:


 * Routing Identifier (RID),


 * Access Identifier (AID),


 * Connecting Identifier (CID),


 * Service Identifier(SID).

The "three mappings" include:


 * the resolution mapping of SID,


 * the resolution mapping of CID,


 * the resolution mapping of AID.

The infrastructure layer introduces virtual access component (VAC) and virtual backbone component (VBC). VAC achieves the pervasive access by introducing the concept and scheme of AID. Users could communicate with each other at anytime, anywhere and share the most popular services. VBC uses RID to solve the problems of routing and management of location. In contrast, VBC realizes the resource configuration and coordination at the core networks by introducing the concept and scheme of RID. The resolution mapping of AID achieves the mapping between AIDs and RIDs and supports four operation modes: one-to-one, one-to-many, many-to-one and many-to-many.

User locates in the VAC, in which the AID represents the identity information of terminal or subnet, being only used in VAC. While the RID in the VBC represents the location information of terminal or subnet, being only used in VBC. The access router realizes the mapping from AID to RID through resolution mapping of AID. By this way, we could implement the separation between indentity information and location information.

The pervasive service layer introduces virtual service component (VSC) and virtual connect component(VCC). VSC is the foundation for realizing the pervasive service. It not only provides the universal description and presentation of various services, but also provides a system to control and manage the services. By introducing SID, various services could be sorted and described in a unified way. The resolution mapping of SID connects the VSC and VCC together and establishes communication tunnels for many kinds of services. VCC introduces the CID as identification of both connection and user. A CID identifies a connection which is used to achieve a service. If a user changes his location,the system only updates the mapping between AID and RID, but the CID and AID both remain invariant. And only the users authorized by the CID could use this connection, other users could not use the connection of this CID unless pass the authentication of this CID. Therefore, VCC could efficiently support mobility and security. In pervasive service layer, the resolution mapping of SID and CID are also employed. Similar with the resolution mapping of AID, both of them support four operation modes: one-to-one, one-to-many, many-to-one and many-to-many.

Identifiers of Identifier Network
Identifier Network employs four different kinds of Identifiers, namely Access Identifiers (AID), Routing Identifiers (RID), Connecting Identifiers (CID) and Service Identifier (SID).

Connecting Identifiers (CID) are used to uniquely identify connections.

Suppose $$Z$$ is the set of routing identifier, theoretically $$Z=\left\{z_1,z_2\cdots z_m\cdots\right\},m\in\left\{1,2,3\cdots\right\}$$, is an infinite set. The routing identifier $$z_i$$ contains the properties of the connection, such as quality of service, bandwidth and so on. The generation of the ith connection can be realized using the following formula:

$$z_i\triangleq \phi \left[ a_1 (i),a_2 (i), \cdots,a_k (i) \right]$$

where $$a_k (i)$$represents kth property of ith connection; $$\phi\left( \bullet \right)$$represents the generation function. The independent variable is the property of the ith connection and the dependent variable is the connection identifier of the ith connection.

Routing Identifiers (RID) are used to route data packets in the core network. One part of RIDs are used for interconnection and interworking between core network equipment, and the other part of them are used for locating, addressing and data forwarding in core network.

Suppose $$Z$$ is the set of switching-routing identifier. Theoretically, $$Z=\left\{z_1,z_2\cdots z_m\cdots\right\},m\in\left\{1,2,3\cdots\right\}$$ is an infinite set. Different from $$X$$, which is the set of Access Identifier, all the elements of $$Z$$ have the same characteristics.

The number of $$Z$$'s subsets is finite. And the $$Z$$'s subsets, $$Z_1,\cdots, Z_m$$, represent different types of switching and routing, such as unicast, multicast and anycast. The set $$S=\left\{Z_1,Z_2\cdots Z_n\cdots\right\}$$ is a partitioning of the set $$Z$$, that is

$$Z_i\subset Z, Z_i\neq \approxeq, and \bigcup_{i=1}^{n} Z_i =Z, Z_i \cap Z_j = \approxeq, (i,j = 1,2,\cdots, n, i\neq j)$$.

The set $$Z_i$$ can be defined as:

$$ Z_i =\left\{ z \mid \varphi _i \left[ p_1 (z),p_2(z),\cdots ,p_k (z) \right] \right\}$$

where $$k\in \left\{ 1,2,\cdots,p_k (z)\right\}$$,$$p_k(z)$$ indicates that the element z should satisfy the kth property, $$\phi\left( \bullet \right)$$ is the caculation formula of routing identifier.

Access Identifiers (AID) are used to identify elements in the edge network such as hosts and routers.

Suppose $$X$$ is a set of access identifiers. Theoretically, $$X=\left\{x_1,x_2\cdots x_i\cdots\right\}$$, where the element $$X_i,i\in\left\{1,2,3\cdots\right\}$$ is also a set. Each element of X represents a type of a terminal, which has its own characteristics different from the other elements.

The AID of the nth terminal which belongs to the ith class can be expressed as:

$$X_i (n)\triangleq \phi _i \left[ a_1 (n),a_2 (n), \cdots,a_k (n) \right]$$

where $$k\in\left\{1,2,3\cdots\right\}$$, $$a_k (n)$$ represent the kth property of the nth terminal. $$\phi _i$$ is the Access Identifier caculation function of the ith terminal. The independent variable is the property of the nth terminal and the dependent variable is the connection identifier of the nth terminal.

Service Identifier (SID) is used to describe all types of data/service in a unified manner.

Identifier network replaces current service identity such as domain name with SID to achieve unified naming for all kind of services. Service identifier is a flat identifier with long lifetime, self-certification, fixed-length and global unique. Moreover, it is independent of the user's position. The flat name structure is non-hierarchical. It does not depend on other structures, which can completely separate the identity and location.

The flat SID structure can be obtained through hashing the service keywords and content. The common hash algorithms are MD-5 and SHA-1. Hash values are 128 and 160 bits, respectively. Considering the large number of future service, we can use SHA-1 algorithm. The SID length is 160 bits. SID is a unique permanent service identifier for routing. However, the SID posesses flat and semanticless structure, which makes it unable to describe a variety of services and to provide flexible source mapping. Thus, assigning SIDs for services is not enough. It is also necessary to design a service description structure to form a complete service identification information unit, making a detailed explanation for service.

Resolution Mappings of Identifier Network
The identifier network employed three kinds of resolution mapping to connect the two layers together. The three resolution mappings are the resolution mapping of SID, the resolution mapping of CID and the resolution mapping of AID, respectively.

In Identifier Network, pervasive service layer employs the resolution mapping of SID and CID. The resolution mapping of SID connects VSC and VCC together. It translates the services into multiple service connections, and each service connection will have its own characteristic according to the specified service. For example, the delay requirements of connections will be different for VoIP and file downloading. Therefore, the pervasive services layer can support various kinds of services. The resolution mapping of SID supports four operation mode, including one-to-one, one-to-many, mant-to-one and many-to-many.

In general, the resolution mapping of SID can be defined as the following fomula:

$$\begin{bmatrix} , & \cdots, & , & \cdots ,  \\ \vdots & \vdots & \vdots \\ , & \cdots, & , & \cdots ,  \\ \vdots & \vdots & \vdots \\ , & \cdots, & , & \cdots ,  \\ \end{bmatrix}$$ $$= \Phi \begin{bmatrix} z_1 (h)_{SID}\\ \vdots\\ z_n (h)_{SID}\\ \vdots & \\ z_N (h)_{SID}\\ \end{bmatrix}$$

where the n represents the service type. m represents the connection type. SID is service identifier. CID is connection identifier. h represents the client. l represents the server. This resolution mapping of SID maps the service $$z_n (h)_{SID}$$ between h and l to a series of connections, that is $$ , \cdots, , \cdots , $$

The resolution mapping of CID connects VCC and VAC together. It translates the service connection into multiple connections in Infrastructure layer. The mapping reflects the idea of “one service to multiple connections” and guarantees the reliability of services.

In general, the resolution mapping of CID can be defined as the following formula:

$$\begin{bmatrix} z_{11}^N (h)_{RID}, & \cdots, & z_{1l}^N (h)_{RID}, & \cdots , z_{1L}^N (h)_{RID} \\ \vdots & \vdots & \vdots \\ z_{m1}^N (h)_{RID}, & \cdots, & z_{ml}^N (h)_{RID}, & \cdots , z_{mL}^N (h)_{RID}\\ \vdots & \vdots & \vdots \\ z_{M1}^N (h)_{RID}, & \cdots, & z_{Ml}^N (h)_{RID}, & \cdots , z_{ML}^N (h)_{RID}\\ \end{bmatrix}$$ $$= \Psi \begin{bmatrix} \\ \vdots\\ \\ \vdots & \\ \\ \end{bmatrix}$$

where the N represents a service, h represents a client, l represents a server, RID represents routing identifier,  and  n and m are the variables. $$\Psi (\bullet)$$ maps n connection identifiers of service N to m routing paths $$z_{ml}^N (h)_{RID}$$ between h server and l client.

The resolution mapping of AID achieves the mapping between AID and RID to connect the VAC and VBC together. The resolution mapping of AID supports four operation modes: one-to-one, one-to-many, mant-to-one and many-to-many.

In general, the resolution mapping of AID can be defined as the following formula:

$$\begin{bmatrix} z_1 (i)_{RID}\\ \vdots\\ z_2 (i)_{RID}\\ \vdots\\ z_T (i)_{RID}\\ \end{bmatrix}$$ $$=\Omega \begin{bmatrix} x_{11} (i)_{AID}, & x_{12} (i)_{AID}, & \cdots, x_{1Q} (i)_{AID} \\ \vdots & \vdots & \vdots \\ x_{21} (i)_{AID}, & x_{22} (i)_{AID}, & \cdots, x_{2Q} (i)_{AID}\\ \vdots & \vdots & \vdots \\ x_{P1} (i)_{AID}, & x_{Pl} (i)_{AID}, & \cdots, x_{PQ} (i)_{AID}\\ \end{bmatrix}$$

where $$z_T (i)_{RID}$$ represents the routing identifier. $$x_{PQ} (i)_{AID}$$ represents the access identifier of terminal system. P represents different terminals and Q represents access location. i represents the type of terminal. $$\Omega (\bullet)$$ is the mapping function, which realizing the mapping between access identifier and routing identifier.

User locates in the VAC, in which the AIDrepresents the identity information of terminal orsubnet, being only used in VAC. While the RIDin the VBC represents the location informationof terminal or subnet, being only used in VBC.The access router realizes the mapping from AIDto RID through resolution mapping of AID. By this way, we could implement the separation between indentify information and location information.