In hetzelfde schuitje Proces automatisering en kantoor automatisering groeien naar elkaar toe
Programma §IT Governance Kantoor automatisering versus Process automatisering §De ketting is zo sterk als z’n zwakste schakel §Een specifiek schakel: de IT-room
In collaboration with
Differences between production and office automation Office Automation
Plant Automation
Transactions, batches, file handling
Real time, “Fire & Forget”
Delay and failures are mostly recoverable or can be repeated. Process time is 1 day / 1 week
Delays and distortions are NOT recoverable
Failures
The cost of failures can be calculated and alternative plans (other locations) can be implemented
The costs of failures can NOT be calculated. There is no limit to the costs of a failure
System
Central computer system with large files systems or databases.
Full distributed system, many small systems, many different vendors.
Availability
Office hours 5*12 hours
Production hours 7 * 24 hours
Calculated availability required
High available required
Relation with user Innovation
Many users for one system (mail, SAP)
Limited users for a system
New releases e.g. once a year. by suppliers,
Iterative process with many small changes anywhere in the process
Relation with vendors
There is one supplier / main contractor
There are many suppliers / contractors, requires management to keep technical quality
Development and changes
Off line tests possible, there is a time for evaluation. Repeated configurations.
By interaction with production hardly time for testing. One – off configurations.
Process
In collaboration with
Process time is 1sec / 1 minute
Kantoor versus Proces automatisering
Kantoor automatisering
Proces automatisering
De tijd die in de SLA hoort te staan (inspanningsverplichting)
Kosten per incident (€)
Acceptabele foutgrens
Acceptabele foutgrens
Detection
De tijd die in de SLA zou moeten staan (prestatieverplichting)
Kosten per incident (€)
Register and start to analyze
1
2
doorlooptijd
3
Solve
4
5
Detection, register and resolve
0. 5
1
doorlooptijd
1. 5
2
De tijd die in een proces omgeving beschikbaar is voor het oplossen van een probleem is zeer kort. Om toch de gewenste beschikbaarheid van een systeem te hebben moet er dus naar andere manieren worden gezocht: Over-dimensioneren, hoge kwaliteit van producten, constante verbetering van de detectie methodes, goede training van procedures, goede en werkbare documenten, zeer strak configuratie management, leveranciers verplichten om pro-actief mee te werken aan verbetertrajecten.
De kosten van een verstoring in een kantoor automatiseringsomgeving zijn redelijk beheersbaar en niet In collaboration with bijzonder tijdkritisch in tegenstelling tot die in een proces omgeving
Application Structure modern production systems Black Box Approach
Future Approach ICT Infrastructure
HMI Presentation Alarming & Events Application Software Operating System Subsystems
Login Security Hardware Monitoring Network Security Backup & Recovery
Central Event Database
Web Interfaces
Directory Services / Single Sign On
HMI Presentation
Login Security
Alarming & Events
Monitoring
Application Software
Network Security
Operating System
Backup & Recovery
Subsystems
Disk Storage
OPC and Other Standards
Central Storage Management
Enterprise Monitoring Tools
Disk Storage
Central Managed Secure OS Scope DCS & SCaDA vendors SCaDA/DCS components Scope ICT
In collaboration with
Enterprise Security Approach Enterprise Backup Solutions
IT Governance in office automation and process automation
Planning /continuity
System Management
Application Administration
IT Policy / Standards.
Automated System Operations (reporting)
User administration
Capacity Planning
Production event Management
Security Admin
Configuration Mgmt
Performance Management
Change Mgmt
Disaster Recovery
Backup and recovery Mgmt
Problem Mgmt
IT systems Technical Application Databases SQL Operating System W2000 Network (Cisco) In collaboration with
Production systems
Supporting systems
MES interfaces (application server) System Management Backup tooling Tools (what’s up gold +(Veritas Netbackup) diafaan)
Databases SCaDA application PLC systems (Siemens) Industrial Network
De ketting is zo sterk als zijn zwakste schakel De verschillende kritische fysieke elementen in de ICT
Infrastructuur – algemene eisen Availability, MTBF The MTBF will increase with: • reliable hardware and software • good change process • good monitoring process • good maintenance
Repairtime, MTTR (detection time, response time, resolve time) The MTTR will increase with: • easy replaceable hardware (COTS) • a good backup / restore process • good monitoring tools • Skilled/trained maintenance personel • simplicity of the system.
Deterministic behaviour of the application. The behaviour will improve with: • QoS for networks • constant monitoring of… − − − −
server load network load Powerload HVAC
In collaboration with
IT-room Een schakel in de ketting
In collaboration with
De uitdaging § Flexibiliteit • Wat zullen de eisen voor de ruimte de komende 10 jaar zijn? − Meer servers? Of meer virtuele servers? − Hogere computer dichtheid? Hot spots − Fusie opkomst? − Outsourcing aan de orde? − SOX-II compliancy?
§ Hoge(re) beschikbaarheid – beperken risico’s • Minder afhankelijk zijn van menselijk handelen • Minimaal N+1 oplossingen • Monitoring van alle vitale componenten − Stroom − Koeling − Blussystemen − Servers − Toegang − Infrastructuur
§ Lage onderhoudskosten / beheerkosten • Noodzakelijke wijzigingen moeten weinig (tijd) kosten • Noodzakelijke wijzigingen moeten weinige risico met zich meebrengen (minimaal N+1)
In collaboration with
De physieke bedreigingen § Brand (EN 1047-2) • Van binnen uit • Van buiten af
§ Inbraak § (Constante temperatuur • IT ruimte is een dark-room • Zeer beperkte toegang (zoveel mogelijk op afstand werken)
§ Continuïteit stroomvoorziening • Netstroom • Noodstroom aggregaat • Onderhoud stroom voorziening • Scheiding van • Noodzakelijke wijzigingen moeten weinige risico met zich meebrengen (minimaal N+1)
In collaboration with
New IT room Expandable / Flexible
§
The IT room (ECB-S certified safe) is capable to contain a total of 33 server racks but can easily be expanded. The 1st floor in the building is prepared to support the weight.
§
The current installation, designed to absorb 11 kW of warmth per rack, can easily and with very little cost be modified to increase the absorption to 25 kW per rack needed when starting with blade servers.
§
Chillers can easily added on top of the roof to improve cooling capacity
§
Total power capacity for the IT room (including the chillers on the roof) is 675kW
In collaboration with
s rverrs e s eesseerv0e tim s d a l BBlad es 220 timeooliningg irires and ccool u q e rre qu w err and ragee e g po e tthhee poitwy ooffaavve ra c y ccaappaaciet nteerrss c nt ddaattaa ce
New IT room Blade server prepared
No w… ionaal l Now… it ion 50 W r TTraadd––it - 8kW 50 W 8kW 6kW 6kW W 00 W 4455 m2 per per rrmm22 ppeem2
11 11
Numbers
3.500.000 3.000.000 2.500.000 2.000.000 1.500.000 1.000.000 500.000
20 times more power needed than conventional data centres!
2000
2001
Blade servers Rack-optimized servers Tower uitvoering
2002
2003
2004
2005 Bron: HP
The specification of operating power density and cooling capacity for In collaboration with data rooms and server rooms is a growing challenge for IT
New IT room Power redundancy (1) NSA NSA
MER – 350kVA? MER – 350kVA?
ATS
Mobile Mobile NSA NSA
80kW
Max. 1600kVA
NSA NSA 1000kVA
1000kVA Total K building Total K building
Nuon
redundant redundant core core
In collaboration with
675kW
New IT room Cooling redundancy NSA MER
Chillers on top of building 40kW 60kW
NSA
Total K building
Mobile NSA
120kW
60kW 120kW 120kW 120kW 120kW 60kW 120kW 120kW 60kW
redundant redundant core core
In collaboration with
40kW 60kW
40kW 60kW
New IT room overview
What
stakeholders
§Expandable / Flexible
IT
§Low TCO
IT
§Aligned with corporate requirements
§High availability
usergroup
§Within budget
project
§EMI protected
M&F department
§Burglar proof
In collaboration with
achieved
Much more to tell…
Vragen?
In collaboration with
