SYSTEM ADMINISTRATION System performance Automating administrative tasks

KENAPA PERFORMANSI OPERASIONAL JARINGAN /SISTIM MERUPAKAN JAMINAN LAYANAN BAGI PENGGUNA UNTUK MENGETAHUI PERFORMANSI SISTIM, DIPERLUKAN MONITORING YANG DAPAT DIGUNAKAN UNTUK MENGANALISA TARGET PERFORMANSI TERCAPAI. SA MEMERLUKAN TOOLS DAN PROSEDUR YANG MEMADAI

JAMINAN SUATU LAYANAN/JASA QoS (JENIS LAYANAN, TINGKAT KEANDALAN DSB) PENGENDALIAN MANEJEMEN (STRUKTUR ORGANISASI, PEMBAGIAN FUNGSI, DSB) SISTIM DOKUMENTASI DLL

MAPPING OS KERNEL MANAGING RESOURCES JARINGAN KOMPUTER INFRASTRU KTUR IT PROTOKOL ROUTING LAYANAN JARINGAN KEAMANAN JARINGAN MANAGING USER & FILES BACK UP & RESTORE MONITORING / AUTOMATING ADMINISTRATIVE TASKS PERFORMANSI SISTIM

System performance ESSENTIAL SYSTEM ADMINISTRATION, AELEEN FRISCH, ch 14 automating adminstrative tasks, ch 14 Managing system resources The Practice of System and Network Administration, Second Edition, Thomas A. Limoncelli, Christina J. Hogan, Strata R. Chalup,

ELEMENTS OF SYSTEM PERFORMANCE (esa, ch 15) CPU (PROCESS PRIORITIES, BATCH QUEUES, SCHEDULER PARAMETER0 MEMORY DISK I/O (FILESYSTEM, I/O RELATED PARAMETERS) NETWORK I/O

PARAMETER PADA JARINGAN SALAH SATU CARA MENGETAHUI indikator kinerja adalah dengan aplikasi SNMP (Simple Network Management Protocol), yang didalamnya terdapat MIB (Management Information base) yaitu struktur database variabel elemen jaringan yang dikelola yang dikelompokkan berdasarkan parameter layanan dan parameter efisiensi. 1.Parameter Layanan 2.Parameter Efisiensi

PARAMETER PADA JARINGAN A. Parameter Layanan Merupakan suatu ukuran yang berorientasi pada pelayanan jaringan dan lebih mempertimbangkan minat pemakai. Parameter ini mengontrol dan merencanakan ketersediaan jaringan yang terdiri atas : 1.Parameter Ketersediaan 2.Parameter waktu respons 3.Parameter keandalan

Parameter Ketersediaan Availibilitas fungsi jaringan adalah presentasi waktu pengguna terhadap akses total layanan jaringan yang tersedia baginya, yang amat tergantung pada keandalan komponen secara teknis. Parameter ini tidak diukur secra langsung, tapi dihitung menggunakan data indikator kinerja (terplan, 1987) yang terdiri dari : – Overall network availability, ketersediaan jaringan secara menyeluruh, diambil dengan mengumpulkan data dari beberapa simpul jaringan yang penting. – Line availability, ketersediaan jalur jaringan, mengukur ketersediaan simpul menerima dan meneruskan paket/frame informasi dari simpul ke simpul dan menginformasikan paketpaket yang diabaikan karena keterbatasan sumber daya (In/OutDiscard) selama pemantauan (SysUpTime). – Customer level availability, ketersediaan pada level user, diukur dari data pemakaian terminal atau simpul terminal.

Parameter waktu respon Indikator waktu respon (terplan, 1987) terdiri dari : – Network delay, penundaan pada jaringan. – Host delay, penundaan pada host dan simpul jaringan, menetapkan lama penundaan waktu dalam komputer pusat untuk suatu paket sampai ke komputer tujuan. – Waktu respon rata-rata – Waktu respon maksimal, menetapkan NILAI maksimum lamanya suatu paket/datagram diizinkan dalam suatu jaringan. – Waktu respon minimal

Parameter keandalan Indikator keandalan (reliabilitas) terdiri dari : – Jumlah failure pada elemen jaringan, memberitahukan suatu gateway tertutup untuk dilewati.. – Daftar tindakan pada kesalahan yang paling sering terjadi. – Jumlah pesan-pesan yang hilang. – Jumlah pesan yang harus diduplikasi. – Jumlah pesan yang tiba, tetapi tidak disampaikan. – Jumlah pesan yang menyatakan kiriman telah diterima. – Jumlah transmisi ulang. – Jumlah time out, waktu yang tidak terpakai karena idle. – Jumlah transmisi yang tidak lengkap.

PARAMETER PADA JARINGAN B. Parameter Efisiensi Merupakan ukuran kinerja yang mementingkan bagaimana informasi bekerja secara efisien dan ukuran daya kerjanya. Daya kerja (throughput) dalam bit per detik, didefinisikan sebagai rata-rata lewatnya bit data pada simpul jaringan tertentu per satuan waktu. Pada jaringan kondisi tetap, kecepatan masuk dan keluarnya paket adalah sama, maka daya kerja adalah harga rata-rata bit per detik tiap memasuki atau meninggalkan jaringan. Indikator parameter daya kerja menurut terplan (1987) terdiri dari : – Transmit, terdiri dari : jumlah transaksi, jumlah paket, jumlah pesan, jumlah karakter, pesan terpanjang dan rata-rata panjang pesan yang dipancarkan. – Receive, terdiri dari : jumlah transaksi, jumlah pesan, jumlah paket, jumlah karakter pesan terpanjang dan rata-rata panjang pesan yang diterima. – Polling, terdiri dari : jumlah pool positif, jumlah pool negatif dan jumlah pool penundaan. – Utilization, terdiri dari : utilisasi pengendalian komunikasi (pemanfaatan protokol IP, ICMP, TCP, SNMP), pengendalian cluster (pada protokol IP) dan utilisasi peralatan terminal – link idle, memantau hubungan yang tidak terjadi antara simpul jaringan pada saat tertentu. – link utilization, memantau hubungan yang terjadi antara simpul jaringan pada saat tertentu. – Contention, memantau terjadinya ‘tabrakan’ pada fungsi perangkat (keras dan lunak) jaringan dan elemen jaringan.

QoS (Quality of Service) Kinerja jaringan diukur dengan metode quality of services (QoS). Pesan yang diharapkan adalah kualitas tinggi dengan menggunakan biaya yang rendah. – QoS jaringan dapat dikarakteristikkan pada 5 pengukuran dasar (Coombs and Coombs, 1998): – Ketersediaan Jaringan (Network availability), rendahnya waktu downtime. – Kinerja yang berhubungan dengan kesalahan (Error performance) – Kehilangan transmisi (kemacetan) dari dua jaringan yang bertukar data. – Waktu yang dibutuhkan untuk membuat koneksi – Kecepatan deteksi kesalahan dan memperbaikinya.

Automating administrative tasks

SERVICE MONITORING Monitoring is an important component of providing a reliable, professional service. The two primary types of monitoring : – real-time monitoring – historical monitoring. Each has a very different purpose. (baca bab As discussed in Section ), monitoring is a basic component of building a service and meeting its expected or required service levels. “If you can’t measure it, you can’t manage it.”

Historical monitoring for recording long-term uptime, usage, and performance statistics. This has two components: collecting the data and viewing the data. The results of historical monitoring are conclusions: “The web service was up percent of the time last year, up from the previous year’s 99.9 percent statistic.” Utilization data is used for capacity planning. For example, you might view a graph of bandwidth utilization gathered for the past year for an Internet connection. The graph might visually depict a growth rate indicating that the pipe will be full in 4 months. Cricket and Orca are commonly used historical monitoring tools.

Real-time monitoring alerts the SA team of a failure as soon as it happens has two components: a monitoring component that notices failures and an alerting component that alerts someone to the failure. There is no point in a system’s knowing that something has gone down unless it alerts someone to the problem. The goal is for the SA team to notice outages before customers do. This results in shorter outages and problems being fixed before customers notice, along with building the team’s reputation for maintaining high-quality service. Nagios and Big Brother are commonly used real-time monitoring systems. Salah satu tools untuk monitoring : SNMP

SNMP SNMP stands for Simple Network Monitoring Protocol. Nobody is sure whether the simple refers to networks or to protocol. Problems with SNMP make it difficult to use on larger-than-simple networks. Although it attempted to be simple, the protocol itself is rather complex. In SNMP’s most basic form, a packet is sent to a network device, such as a router, with a question called a GET. For example, one might ask, “What is the value of IF-MIB::ifOutOctets.1?” That variable is in the group of interfacerelated (IF) variables, the one that records how many bytes (octets) were sent out of the interface (ifOutOctets), on interface number 1. The router replies with a packet containing the value. There are variables for just about everything and every kind of technology. A group of related variables is called a MIB. There are standard MIBs for Ethernet devices, DSL devices, ATM devices, SONET devices, T1/E1 devices, and even non-network technologies: disks, printers, CPUs, processes, and so on.

Simple Network Management Protocol (SNMP) SNMP has three components: The SNMP Agent The Managed Device The Network Management System (NMS) The SNMP agent is software that runs on all network devices – computers, routers, printers, and switches, to name a few. The agent is responsible for ‘monitoring’ the device, and if there is a problem, relaying that information to the NMS. (continued)

SNMP (continued) The Managed Device is any host device in which an agent has been installed. The NMS is responsible for collecting the information the agents send to it. The agent uses a management information base (MIB) to compare the current readings to the standard readings in the MIB. If there is an aberration, the agent will notify the NMS. (continued)

SNMP (continued) The agent knows three types of “commands” or utilities: – GET tells the agent to find some piece of information about the managed device and forward it back to the NMS. – SET tells the agent that the NMS wants the agent to change a configuration setting. – A TRAP is the notification to the NMS when there is a problem at the managed device. All agents belong to a community. An agent will only report to an NMS that is part of its community. This is considered a low-level security feature.

Alerting Mechanism The monitoring system has an alerting mechanism to make you aware that something requires human attention. There is no point in software knowing that something has failed or is overloaded unless it tells a human about it or does something about it and makes a note of it for a human to look at later

Escalation Another component of the policy and procedures that your monitoring and alerting system should implement is the escalation policy describing how long each problem should be permitted to persist before it is escalated to another person, typically a manager. The escalation policy ensures that even if the person receiving the alert is on vacation or doesn’t respond, the issue will be passed on to someone else. The escalation policy needs to describe various escalation paths for various categories of alerts.

Active Monitoring Systems An active monitoring system processes the problems it detects and actively fixes the ones that it knows how to deal with. For example, an active monitoring system might reset a modem port that it detects is in a strange state or might remove a modem from a modem pool if it could not be fixed by a reset. Active monitoring systems can be useful up to a point. Although they respond more quickly than a human can, they have limitations. In general, an active monitoring system can implement only a temporary fix. The system won’t detect and permanently fix the root of a problem, which is what really needs to happen (see Chapter 16)