Technology Support for Smart Computational Grid (SCG) in Future Wireless Systems

The communication grid is an emerging infrastructure that will fundamentally change the way we think – and use – computing. Communication grid is used by analogy with the electric power grid, as an infrastructure that provides pervasive access to electricity. Like the computer and a small number of other advances had a dramatic impact on human capabilities and utilization of the scarce frequency resource, it is believed that allowing all components of our information technology infrastructure – computational capabilities, databases, sensors, and people – to be shared flexibly as true collaborative tools, the Smart Computational Grid (SCG) will have a similar transforming effect, allowing new classes of applications to emerge. Improvement in spectrum usage achieved by CR technology and associated collaborative computations on white spaces. Every spectrum hole includes several channels. Implementing collaborative algorithms over the channels help track changing environment parameters and managed to achieve smooth handoffs, which is maximized to offer services ranging from low-data-rate non-reatime applications to high-speed real-time multimedia applications. With virtualization, the heterogeneous wireless access networks complement each other to deliver maximal signal strength, wider coverage, efficient data rate, low latency and reduced loss rate.

intelligently detect communication channels where all network entities have the same right to access the spectrum. Multiple access is allowed for coexistence and communication using the same portion of spectrum. Fig. 1 Multiple Access Points (APs) centrally architecture Similar to a grid network, the CR technology enables transmission parameters of many wireless nodes within its network to be changed to communicate while avoiding interference with original users. Therefore, configuration in cognitive networks became full reconfigurable system that automatically varies its communication variables with network and user demands (Ghosh, Das and Chatterjee, 2014). Technically, each communicating terminal is expected to be characterized and equipped with varied technologies and standards (IEEE 802.11*,802.15*,802.16*,802.22 and xG) and the interfaces includes UMTS, W-LAN/Wi-Fi, Wi-MAX, WSN, WRAN and others. Cognitive Radio Networking was thought of as an ideal goal of providing highly reliable communications whenever and wherever needed for efficient radio spectrum utilization. Hence the main goal of spectrum sensing to determine spectrum status and licensed user's activity is achievable with co-operative sensing as a form of distributed sensing. Provisioning for co-existence of heterogeneous devices operating in the spectrum holes is a drive for need to collaborate the sensing required to detect PU signals. Also, limited availability of spectrum and the non-efficient use of existing RF resources necessitate a new communication paradigm to exploit wireless spectrum opportunistically and even with greater efficiency (Mauri et al, 2014).
In view of the above, this paper is geared towards the objective of:  modeling a framework for computation;  collaboration for mutual parameter estimation; and  intersystem handoff procedure to provide ubiquitous coverage in future systems

Review on grid computing
Spectrum management functionalities, which are in distributed co-ordination of CRN includes spectrum sensing, spectrum sharing, spectrum decision and spectrum mobility (Salem, AbdElKadir, Ramadan and Abdel -Mageed, 2014). The three basic types of spectrum sensing techniques for detecting PU licensed spectrum band outlined in Mauri et al (2014), therefore include co-operative, transmitter and interference-based spectrum sensing techniques. Cooperative detection allows multiple CRs to work together and supply required information to detect a PU. Spatial diversity intrinsic to multiuser network is exploited in distributed or centralized structure. Distributed approach enables CRs exchange spectrum observations among themselves while each develops a spectrum occupancy map based on signal information detected. This approach mitigates multipath fading and shadowing RF problems to increase probability of PU detection and also help combat common hidden node problem peculiar to ad hoc wireless networks. The optimal detector in stationary Gaussian noise of communication channel is the matched filter, which maximizes received SNR of transmitted signals (frames). Matched filter technique of energy detection works by correlating known signal template with unknown signal and detecting presence of template in the unknown signal. Less time (frame duration) is required to achieve high processing gain due to coherency. Unlike other signaling techniques, (cyclo-stationary and interference-based) with long sensing time and complex computational tasks, matched filter detection only requires a prior knowledge of PU signal. This is readily made available in real-world situation as CR user monitors radio spectrum periodically to opportunistically communicate over spectrum holes.
Computer Engineering and Intelligent Systems www.iiste.org ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol.10, No.7, 2019 Spectrum Sharing and DSA With spectrum sensing as first step in realizing cognitive radio communications, overview of MAC layer sensing issues and MAC protocol performance were outlined in Bayhan and Alagoz (2012). An enhancement for analyzed MAC scheme was also proposed as basis for a channel to forming cluster of sub-channels. The clustered approach enables the secondary nodes clustered into groups and each SU contends with the elements in the same cluster. This arrangement form set of sub-logical channels instead of a single logical channel.
PHY layer sensing accuracy highly depends on the sensing duration and in CR system, authorized or licensed user has the priority to occupy the frequency band, while the cognitive users (SUs) may only use the spectrum opportunistically. CRN enabled SUs utilize spectrum holes detected in licensed bands for their communication needs. Spectrum holes are free of PU existence or are partially occupied by low-power interferes.
CR operating parameters includes normalized transmission power, modulation scheme, transmission bit rate and centre frequency [sub-carrier bandwidth]. These are adjusted in order to operate in the dynamic spectrum usage required by next-generation heterogeneous wireless environments (Raman and Singh, 2014). Since static allocation of the frequency spectrum does not meet the needs of growing wireless technology, CR technologies offer efficient spectrum sharing through dynamic access (Chowdhury et al, 2012) and so DSA technique is capable of addressing spectrum scarcity problems. DSA as new spectrum sharing paradigm, allows SUs access to abundant spectrum holes in licensed bands (Song, Xin, Zhao and Cheng, 2012).
The reliability of cognitive operation depends on effective spectrum sensing as remarked in Raman and Singh (2014). Song et al earlier observed the need for new DSA model where licensed users would co-operate to provide much more flexible spectrum sharing. Also, to further solve the problem of spectrum scarcity, Ejaz et al (2012) described energy detection scheme as simple strategy with poor performance under low SNR.

Threshold Setting in Signal Detection
PHY-layer sensing adapts modulation schemes and parameters to measure and detect the primary users' signals on different channels. PHY-layer detection methods -energy detection, matched filter and feature detection are commonly used with feature detection being a good candidate for detecting PU's presence on channel as energy detection method. Detection results are always given in the following three possibilities of: channel is idle; channel is occupied by its primary users, but secondary users are allowed to transmit their packets with some power constraints; or channel is not available to secondary users at all. On the other hand, the MAC-layer sensing determines when and how a SU node sense and which channel(s) it senses (Firooz, Chen and Liu, 2013).
Two major threshold setting scheme in literature include probability of detection ( d P ) and probability of false alarm ( f P ), which is usually fixed to a small value (minimized) while d P is allowed to a big value (maximized). To maximize d P requires a constant false alarm rate (CFAR) principle. Similarly, to minimize f P while guaranteeing a non-interference probability, miss detection probability ( m P ) is set to a minimum and d P to a high value, the approach is called constant detection rate (CDR) principle.

Detection Modeling
Generally, received signal y over CR model is expressed as z hx y    Bayhan and Alagoz (2012) and Raman and Singh (2014), PHY layer energy detection expression for the signal received by SU is given as Where the number of sample signal used in sensing is unlimited (scalability offered by cognition), energy detector provides for d P and f P simultaneously. The minimum number of sample signal subjected for detection is a therefore function of SNR evaluated and is given as where θ is the transmitted power

Framework for collaborative computation
Each wireless terminal is enabled to coexist with each other while being able to select a good channel. Though, characterized with intra-multiple user co-operation or competition collision challenge, the super-layer accurate sensing is guided against interference constraints. As depicted in fig. 2, PHY layer depend on sensing duration of other user detecting available spectrum opportunity locally whereas MAC sensing depends on sensing period and search order, but network layer while the network layer sensing is depicted by the SU co-operation.
Communicating channels are sensed using an estimation of channel occupancy probabilities as ordered by Latin square. To enable the multiple users maintain collision-free sensing, minimal false alarm is desired and each node is equipped with a cognitive radio to offer cogntion and reconfiguration techniques.  Fig.3 Three CRs sensing three channels within same sensing duration In scenario (a) and (c), CR-1 and CR-2 chose the same sensing order but collision was avoided by CR-1 in (a) by finding channel 1 free in step 1 while CR-2 generate a false alarm in that step. CR-1 and CR-3 find channels 3 and 1 free in steps 4 and 5. In scenario (b), only CR-1 find channel 1 free in step 1 because the same sensing order was not used by CR-1 and CR-2. CR-2 find channel 2 free in step 4 while finding channels 1 and 4 CR-3 find channel 2 free in step 3.
In scenario (c), CR-1 and CR-2 both find channel 4 busy in step 4 because of using same sensing order although this is resolved by CR-3 detecting non-existence of PU signal (H0) in channel 4 (step 2) while CR-1 and CR-2 still find channel 5 busy. The same sensing order for two or more CRs is prevented to arrive at collisionfree detection of spectrum holes. From the framework, entire frequency band (licensed) have its allocated channels occupied (busy) mostly at every step servicing either PU or SU.

Implementing Slotted Rendezvous for Collaboration
Major objective of many CRs in the scenario is to offer intelligence and fair spectrum sharing to achieve collaboration and enhance the performance of the SCG while implementing radio rendezvous processes. As channels are sensed with highest probability being idle channel sensed first (CR-1 sensing channel 1,2,3,4,5 in scenario (b) while CR-2 in step 1 senses channel 3 first being the channel with highest probability of being idle since CR-1 had detected PU signal on it in step 3. With CR-1 and CR-2 sensing with the same order in (a) and (c), a rendezvous of obtained.
An ability of two or more radios meeting to establish a link on same channel provides for bootstrapped communication requirement of multi-channel systems. With this arrangement, SU is adversely affected in its sensing as the link maintains channel availability (DaSilva, 2012). Need for increased co-ordination between CR nodes is desired for detected available frequency bands (holes) to be joined together to from sub-channel logical band for SU access. A proactive plan to combat changing environmental conditions requires major task of detection probability accuracy.
While targeting how to maximize d P or minimize f P , an optimized CFAR and CDR is evidently provided by the slotted rendezvous, which dynamically allows all channels {1,2,…,N} to be visited by radios in random, then in prescribed order scenario (b) and (c). Here, collaborative sensing by CRs increases system throughput since time to rendezvous (TTR) increases with N. Although, slotted rendezvous is equally likely to happen in any of the N channels, real-time sensing in search of white spaces maximizes TTR. This is a function of SNR of the signal samples. CRs multiplicatively decrease the probability of using same sensing order for two or more CRs. Frame duration comprises time slot for sensing, searching and transmission  (2012), each access frame has a fixed duration divided into two partssensing slot and packet transmission slot. The sensing slot comprised of sensing time and channel search. Therefore, frame duration is a combination of sensing time (Ts), channel search (Tc) and Transmission (Tt), the efficiency of spectrum usage for any typical sensing duration is expressed as

Discussion
To have a constant desired d P in dynamic scenario, number of samples can be varied for lower noise levels. The theoretical analysis shows that coherent processing can turn low SNR into high SNR regime so that, given enough samples, arbitrary weak signals can be detected. Scenarios (a) and (c) demonstrates performance of collaborative spectrum sensing in sequential order while scenario (b) showed sensing in random and then sequential order. When environment parameters change, SNR values changes f P , which in effect changes for desired d P .
The dynamic environment evident in SCG, higher f P is obtainable for small changes in d P , which is one of the major metrics of measuring performance. Different values of d P can be obtained and evaluated for different SNR and channel selection based on incumbent appearance expectations in further research. Traditional channel selection mechanism could also be enhanced as well.

Conclusion
Consumption power, throughput, error rate and interference performance objective metrics of SCG has greater dependence on minimization of power consumption, which is directly placed on transmission power (P) in agreement with Chen and Wyglinski (2009). Improved sensing capability is devised to track all changing environment condition as implemented in the computation. This connote that minimum detectable signal level and sensing time is required to achieve desired d P and f P . Higher d P implies reduced f P , and this increases CR-SU exploitation of the unused spectrum adaptively in the radio environment. Improvement offered by number of collaborative access point (AP) radios and threshold rule offer remarkable improvement in spectrum usage efficiency. This is further guaranteed with multiple cognitive terminal cooperating to maximize frame duration than individual terminal computing. Real-time sensing is offered by the framework in search of white spaces. This enable the SCG benefits tremendously from a proactive plan especially with changing noise variance along communication channels.