Page 964 - Cloud computing: From paradigm to operation
P. 964

6                                                    Monitoring


            See Table 7-8.


                               Table 7-8 – Parameters collected for the monitoring environment

                        Metric name                            Description                       Unit

             temperature                       Current temperature of server room                 o C
             voltage                           Current voltage of electricity supply               V
             electric current                  Electric current of electricity supply              A
                                               Output voltage of an uninterruptible power
             UPS voltage                                                                           V
                                               system (UPS) battery
             UPS output electric current       Output electric current of UPS battery              A
             UPS charging electric current     Electric current of charging                        A
             humidity                          Relative humidity of server room                    %
             fan rotations                     Fan rotations per minute                       rotations/min
             power of fan                      Current power of fan consumption                   W

            7.2     Service layer parameters

            7.2.1   General

            One of the most important areas for provider and consumer is service performance and availability when it
            comes to cloud computing. Cloud service provider and consumer need to get an entire view of the health of
            service. A lot of decision making and SLA determination are driven by service performance and availability.
            The  monitoring  system  should  report  the  service  performance  and  availability  parameters  to  identify
            whether the QoS specified in the SLA is fulfilled.

            Different  cloud  services  can  be  offered  with  different  terminologies,  specifications  and  features.  Cloud
            services  can  achieve  different  levels  of  performance  under  various  workloads  generated  by  diverse
            applications. For example, unlike computation and communication-intensive applications, performance of
            data-intensive applications typically will be strongly affected by I/O performance and storage access in a
            cloud infrastructure. The monitoring parameters of typical services are included in this clause. Others are for
            further study.
            See Table 7-9.


                             Table 7-9 – Performance metrics for monitoring general cloud services

                    Category             Metric name                           Description

                                    MTBF                    Mean time between failures
             Availability
                                    MTTR                    Mean time to repair
                                    Response time           Response time for composite or atomic service
             Performance                                    Number of transactions or requests processed per
                                    Throughput
                                                            specified unit of time
                                    Bandwidth               Bandwidth of the connection that supports a service
             Capacity                                       Capacity of a temporary or persistent storage medium,
                                    Storage capacity
                                                            such as RAM, disk or tape











            956
   959   960   961   962   963   964   965   966   967   968   969