Based on the data occupying the buffer at each ONT, the OLT will create a Bandwi

Question

Based on the data occupying the buffer at each ONT, the OLT will create a Bandwidth Map in the form of (ONT_id. Start Time. Stop Time) for each ONT with non-empty buffers. To minimize the gaps between Upstream transmissions, the OLT should schedule the StartTime and StopTime for each ONT so as to minimize the gaps among the received data. This requires the OLT to know the transmission delay between each ONT and the OLT. You are required to write a pseudo code algorithm to create the GPON UpStream Bandwidth Map of a set of ONTs based on their respective transmission delay and bandwidth requirement. The resultant BandWidth Map should not have a wasted bandwidth other than the 1 milli second guard time between ONT transmissions there are 32 ONTs connected to the OLT All ONTs are discovered and operational (no new ONTs) All times are in milli seconds Bandwidth requirements are random ONT transmission delays are random Psuedo code of algorithm with description of every step Flow chart of algorithm An example of a possible resultant Bandwidth Map

Explanation / Answer

DSCA Flowchart The flowchart used for the implementation of the DSCA monitoring algorithm is displayed in Figure 4, explaining the individual conditions for subcarrier utilization as well as defining all parameters [13]. Figure 4: Flowchart of DSCA algorithm with monitoring. The algorithm assumes that for each ONU an electronically imposed equalization delay is applied and, as a result, the OLT can consider them all located at the same distance, similarly to GPONs. The flowchart of the DSCA algorithm consists of the following steps: STEP 1: By the end of each window, the OLT calculates the average subcarriers used per ONU during the preceded monitoring window time. STEP 2: As a result ONUs are partitioned into two groups; the overperforming and underperforming ONUs: • overperforming group: Pre_SC = Used_SC. • underperforming group: Pre_SC < Used_SC. STEP 3: (for underperforming group from STEP 2): The OLT compares Used_SC with SLA_SC using different reference subcarriers to distinguish SLA grades as applied for ONU 2 at Figure 3. The following cases are considered: Calculation of Used_SC Pre_SC > Used_SC Compare Used_SC with SLA_SC Compare Pre_SC with SLA_SC Allocate remaining subcarriers (Assigned_SC) Yes Flag = OFF No Flag = ON Variables Used_SC: Used subcarriers per ONU during Monitoring window SLA_SC: Originally assigned subcarriers Pre_SC: Subcarriers assigned at the previous window Assigned_SC: Subcarriers for the next window STEP 1 STEP 2 STEP 4 STEP 3 STEP 5 STEP 6 Definition and evaluation of algorithms for dynamic bandwidth allocation in ACCORDANCE Accordance_D4.3_WP4_2011_31October_UH_v1.0.docx ACCORDANCE FP7 – ICT– GA 248654 Page: 18 of 67 • If Used_SC is smaller than or equal to SLA_SC, then the OLT defines Assigned_SC based on the Used_SC, subtracts Used_SC from SLA_SC and assigns the difference to the group of “remaining subcarriers”. • If Used_SC is greater than SLA_SC, then the OLT defines Assigned_SC equal to SLA_SC because it is not aware at this stage if there are “remaining subcarriers”. STEP 4: (for overperforming group from STEP 2): The OLT compares Pre_SC with SLA_SC without considering the Used_SC since it has already determined in STEP2 that ONUs require more subcarriers as applied for ONU 1 and 3 at Figure 3. The additional subcarrier allocation is performed as follows: • If Pre_SC is smaller than SLA_SC, then the OLT increments Pre_SC by 1 (Pre_SC + 1) and allocates it to Assigned_SC. Then it subtracts Assigned_SC from SLA_SC and assigns this difference to the “remaining subcarriers”. • If Pre_SC is greater than or equal to SLA_SC, then the OLT defines Assigned_SC equal to SLA_SC because it is unaware at this stage if “remaining subcarriers” are available. STEP 5: After completing STEPs 2, 3 and 4 the OLT gathers the “remaining subcarriers” from the first case of STEP 3 and STEP 4 and distributes them to requesting ONUs based on their SLA priority. STEP 6: Following STEP 5 if there are “remaining subcarriers” the OLT assigns them to ONUs based on SLA priority. 3.1.2. Network Simulation and Perf

Network Simulation and Performance Evaluation The simulation model exhibits an OFDMA-PON composed of one OLT and 32 ONUs, spanning over 40 km. Three SLA types, SLAt , t=0, 1, 2, from high to low superiority have been considered. The number of ONUs in each service level is set to 2, 10 and 20 with the buffer size of each ONU limited to 10 MBytes. The total upstream data capacity is 10 Gbps, arranged in 64 subcarriers of 156.25 Mbps each. In addition, the guaranteed bandwidth for each SLA from high to low is specified at 468.75 Mbps, 312.5 Mbps and 156.25 Mbps respectively. Grant processing and propagation delays are 0.5 µs and 0.5 µs/km respectively. The network traffic is implemented by a Pareto selfsimilar traffic model with a typical Hurst parameter of 0.8 to simulate practical network patterns. The packet size is uniformly generated between 64-1518 Bytes. Definition and evaluation of algorithms for dynamic bandwidth allocation in ACCORDANCE Accordance_D4.3_WP4_2011_31October_UH_v1.0.docx ACCORDANCE FP7 – ICT– GA 248654 Page: 19 of 67 The simulation parameters required for data link layer modeling and protocol evaluation are summarized in Table 1 below alongside, CoS differentiation and traffic generation. Table 1: Example of DSCA design parameters. Parameters description Total network capacity 10 Gbps Number of subcarriers 64 Data rate per subcarriers 156.25 Mbps (10 Gbps / 64) Number of ONUs 32 SLA0 : SLA1 : SLA2 = 2 : 10 : 20 Guaranteed Data Rate per SLA (SLA_SC) SLA0 468.75 Mbps SLA1 312.5 Mbps SLA2 156.25 Mbps Distance between OLT and ONU 40 km Monitoring window time 2.0 ms Grant processing delay 5 µs Propagation delay 5 µs/km ONU offered load 1.0 312.5 Mbps (10 Gbps / 32) Network offered load 1.0 10 Gbps Packet size 64 – 1518 Bytes (Uniformly generated) Traffic generation of CoSs High priority CoS0: 20 % Middle priority CoS1: 40 % Low priority CoS2: 40 % Figure 5 (a) confirms that the end-to-end packet delay of high and middle priority SLAs is less than 1.4 ms, even if the ONU offered load is 1.0. This is because the guaranteed bandwidth of the high and middle SLAs are greater than or equal to the ONU offered load of 1.0 corresponding to 312.5 Mbps. In addition, the DSCA algorithm allocates the remaining subcarriers according to the SLA priority. As expected, the end-to-end pack delay of low priority SLAs dramatically increases at an ONU offered load of 0.7 since DSCA on its own cannot support granularity and further Definition and evaluation of algorithms for dynamic bandwidth allocation in ACCORDANCE Accordance_D4.3_WP4_2011_31October_UH_v1.0.docx ACCORDANCE FP7 – ICT– GA 248654 Page: 20 of 67 subcarrier distribution among low bandwidth ONUs. Figure 5 (b) represents the end-toend packet delay according to three CoS profiles under SLA0 to establish the priority insights for the various service classes accessing the network, at conditions of high bandwidth transfer. It is worth pointing out that the packet delay of the high priority traffic, CoS0 is less than 1.0 ms, a very promising figure considering it is reported at 40 km link spans and the capability such a figure offers to ACCORDANCE to support the stringiest of requirements of its various backhauled technologies. Additionally the observed trace, compared to the remaining classes, displays a trend at the right direction since CoS0 is expected to depart ONU queues with high priority. Similar characteristics were also recorded for SLA1 and SLA2

Related Questions

Navigate

• Browse (All)
• Browse B
• Subjects

---

---

Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.