TIPC: Transparent IPC

 

As now days applications are growing in exponential way. Mobile phones are now only for voice communication. Now days voice communications are secondary for development prospect. Now the more focus is on proving high availability of services. To provide every architecture has now support of distributed environment. So when we talk about distributed systems we are talking about multiple instances of our data. So whenever there is some changes in one instance of data that must be properly synchronized with other replica as well. So for that we have to think about transport protocols like TCP or SCTP. So these protocols have their set of requirements for establishing communication.

 

So when there is internal communication among replicas of data, we found TCP little clumsy. So Ericson has developed and deployed a new solution for their products named Transparent Inter process communication (TIPC). TCP lacks functional addressing and addressing transparency. TCP uses mapping in general statically, so it is very poor approach to use it in dynamic real-time environment. TCP is good enough for large messages but when we have to just synchronize with very short message still we have to follow the whole cycle of TCP protocol standard. So to overcome that cycle TIPC uses some direct communication mechanism having address transparency. So where in case of short message TCP uses minimum 9 packets transfer , TIPC can do that in 1 or 2 packets flow. So if we check for inner node transportation of messages we have 35% better performance over TCP.

 

Assumptions for better implementation of TIPC

1. Most messages cross only one direct hop.
   
2. Transfer time for most messages is short.
   
3. Most messages are passed over intra cluster connections.
   
4. Packet loss rate is normally low; retransmission is infrequent.
   
5. Available bandwidth and memory volume is normally high.
   
6. For all relevant bearers packets are check-summed by hardware.
   
7. The number of inter-communicating nodes is relatively static and limited at any moment in time.
   
8. Security is a less crucial issue in closed clusters than on the Internet.
   

 

Uncanny Valley

After watching a number of Hollywood action movies, like Avatar, Terminator specially science friction ones, If I tell you that I have a android which can sing , dance and perform all steps with real human. Now days Japan has turned into Uncanny valley where they are just about to insert humanoid with human community.

Recently a new member joined the Japanese band AKB48. Pop blogs and a magazine cover story introduced Aimi Eguchi as a sweet 16-year-old from a Tokyo suburb. But there was something strange about this new girl. Her incandescent looks and sterling voice won Aimi instant attention and the center spot in a candy ad. Yet questions arose when she appeared in two video spots. Aimi seemed stiff and awkward – not uncomfortable, more like unnatural.

"There was a weird reaction to it," says Zac Bentz, a writer and reviewer for the online music store HearJapan.com. "Within a week or two, people were already saying, 'Well, this doesn't look right.' "

She always looked straight at the camera, shared a bizarre resemblance to other girls in the band, and had a peculiar beauty. After a few weeks, AKB48 admitted that Aimi was computer generated. They took the nose, eyebrows, hair, and lips of six band mates and digitally stitched them into a new singer.

Future Map

Till we have come to know about robots which can perform specific tasks for which they have trained. The use of Artificial intelligence was for their learning and carve their skills in specific fields. They were better in working but not intelligent as we. I have seen movie robot of Rajnikant where he succeed in emotions mapping in Robot. But if a robo can sing dance and can read our minds with the help of 3D imaging technology and brain mapping. Advance robot can map human mind and can get idea about his/her feelings. He can map his body by 3D imaging technology. All the matter is that we are going to adopt them as we are excited to accept this new generation. Till now Japan is uncanny valley, very soon we could acknowledge that we are also a part of it.

So carve yourself and be ready to accept a humanoid among us without knowing his actual identity. Its surely be adventures but hoping not to be like resident evil series.

Legacy Connectivity (3GPP Rel-4)

There are support for legacy network but first we need to understand Dial-up connections and GPRS. Both basically related to initial development of 3GPP standards (up to Rel-5). 

Specially for mobile user, which are used for providing information on small screen of mobile. So WAP was not fully compatible to public internet so WAP Gateways are used to get data and send request over public internet.

Dial up Connection:

Figure 1: GSM circuit switched access connection

While connecting with dial-up connection we use a E-164 number to dial to operator who connects to a WAP server or third party service provider. In initial deployment we have used RADIUS protocol so when a user dials up E-164 NAS number, user request of NAS connects to gateway, so at gateway there is authorization check is made and if allowed to access public internet, Gateway provide a private IP address to User terminal. By this way  a user can connect to public internet as Gateway publish a Public IP of user for other network and by getting data from that end , it map address of public identity to user private IP Address.

GPRS connectivity:

The GPRS nodes of each operator are interconnected on a private IP network.  The access connection (the technical name is PDP Context) can be thought of as a “flexible tunnel” through the GPRS networks that is established by the GPRS Tunneling Protocol (GTP). The arrangement is shown in Figure 2. The GTP tunnel extends from the SGSN to the GGSN. A mobile may support several PDP Contexts simultaneously.

Figure 2: The GPRS access connection
 

A GPRS user can access the following types of access point:

·        a Virtual Private Network connected to their home mobile operator (APN = <name of VPN>.mnc.mcc.gprs)

·        a third party ISP connected to their home mobile operator  (APN = <name of ISP>.mnc.mcc.gprs)

·        the Internet using either normal Internet protocols or WAP via the visited network’s connection to the Internet (APN = “Internet”).

APN name is an IP address of access point. APN is not URL of information source which user wants to interrogate, it is a access point by which a user can access many information source.

For the tunnel between the SGSN and the GGSN, i.e. the section across the GPRS backbone, there is a tunnel  identifier (TID) distinguishing each user’s tunnel. The tunnel ID relates to the GTP protocol running between the SGSN and the GGSN, and there are IP addresses for the source and destination SGSN/GGSN interfaces at each end of the tunnel. So this approach is known as circuit switching communication.

General Information Elements

Wireless system Analysis is based on two part. Radio Access Network part and Core technologies.

In the Radio Part, the technologies are evolving towards higher efficiency ( i.e. being able to encode/or send more data on the same radio signal ( or the carrier frequency). The movement is from TDM/FDM ( or 2G GSM), CDMA (IS95, CDMA 1x ) to WCDMA ( 3G) to OFDM (4G & LTE) .
In the Core, technologies are evolving from circuit switched to being packet switched. This basically means that instead of reserving a whole circuit for the duration of the call, you are dividing it in to data packets , which are re-assembled at the recieving end. IMS has multiple releases and stages, through which this move is being achieved.

IP Multimedia Service (IMS) is an architectural framework for delivering Internet Protocol (IP) multimedia services. It was originally designed by the wireless standards body 3rd Generation Partnership Project (3GPP), as a part of the vision for evolving mobile networks beyond GSM. Its original formulation (3GPP R5) represented an approach to delivering "Internet services" over GPRS.

Relational Analysis of LTE, IMS, SAE, EPC

LTE (Long Term Evolution) is the 4G wireless access technology from 3GPP like UMTS was the 3G or GPRS the 2.5G. Because it provides IP based only access and a lot of bandwidth 3GPP decided to study what converged architecture will support it. The study was called System Architecture Evolution (SAE).
The Evolved Packet Core is the result of this study. It is the All-IP architecture to which LTE and other 3GPP (UMTS,GPRS) and non-3GPP (WiMAX,HRPD,WLAN etc) access systems connect to. The EPC provides a converged solution for Security, QoS , Mobility and connection to the IP based services (IMS or the Internet).Sometimes the term Evolved Packet System (EPS) is also used but it just refers to the EPC and the LTE access network (E-UTRAN) together.
IMS is on top of the EPC but its just considered one of the possible IP Services layer. There is a data interface (SGi) from the main gateway of the EPC (the PDN-Gw) to IMS and also a signaling/control interface (Rx) from the application function (the P-CSCF in IMS) to the main session controller in charge of the authorization, admission control, resource reservation, QoS (PCRF).

Basics of LTE Understanding

Before we start discuss anything first come to knowledge that IMT has set some requirements that must be fulfilled to qualify for 4G naming.
IMT-Advanced standard requirements

The main purpose of these processing is to provide better speed and reliability on wireless side of mobile communication.

• Peak download rates of 326.4 Mbit/s for 4x4 antennae, and 172.8 Mbit/s for 2x2 antennae (utilizing 20 MHz of spectrum).
• Peak upload rates of 86.4 Mbit/s for every 20 MHz of spectrum using a single antenna.
• Five different terminal classes have been defined from a voice centric class up to a high end terminal that supports the peak data rates. All terminals will be able to process 20 MHz bandwidth.
• At least 200 active users in every 5 MHz cell. (Specifically, 200 active data clients)
• Sub-5 ms latency for small IP packets.
• Increased spectrum flexibility, with supported spectrum slices as small as 1.4 MHz and as large as 20 MHz (W-CDMA requires 5 MHz slices, leading to some problems with roll-outs of the technology in countries where 5 MHz is a commonly allocated amount of spectrum, and is frequently already in use with legacy standards such as 2G GSM and cdmaOne.) Limiting sizes to 5 MHz also limited the amount of bandwidth per handset
• In the 900 MHz frequency band to be used in rural areas, supporting an optimal cell size of 5 km, 30 km sizes with reasonable performance, and up to 100 km cell sizes supported with acceptable performance. In city and urban areas, higher frequency bands (such as 2.6 GHz in EU) are used to support high speed mobile broadband. In this case, cell sizes may be 1 km or even less.
• Good support for mobility. High performance mobile data is possible at speeds of up to 350 km/h, or even up to 500 km/h, depending on the frequency band used.
• Co-existence with legacy standards (users can transparently start a call or transfer of data in an area using an LTE standard, and, should coverage be unavailable, continue the operation without any action on their part using GSM/GPRS or W-CDMA-based UMTS or even 3GPP2 networks such as cdmaOne or CDMA2000)
• Support for MBSFN (Multicast Broadcast Single Frequency Network). This feature can deliver services such as Mobile TV using the LTE infrastructure, and is a competitor for DVB-H-based TV broadcast

E-UTRAN Air Interface

E-UTRAN is the air interface of LTE. Its main features are:

• Peak download rates up to 292 Mbit/s and upload rates up to 71 Mbit/s depending on the user equipment category.
• Low data transfer latencies (sub-5 ms latency for small IP packets in optimal conditions), lower latencies for handover and connection setup time than with previous radio access technologies.
• Support for terminals moving at up to 350 km/h or 500 km/h depending on the frequency band.
• Support for both FDD and TDD duplexes as well as half-duplex FDD with the same radio access technology
• Support for all frequency bands currently used by IMT systems by ITU-R.
• Flexible bandwidth: 1.4 MHz, 3 MHz, 5 MHz, 15 MHz and 20 MHz are standardized.
• Support for cell sizes from tens of metres radius (femto and picocells) up to 100 km radius macrocells
• Simplified architecture: The network side of EUTRAN is composed only by the enodeBs
• Support for inter-operation with other systems (e.g. GSM/EDGE, UMTS, CDMA2000, WiMAX...)
• Packet switched radio interface.