Upload
lamtuong
View
216
Download
2
Embed Size (px)
Citation preview
© Square Mile Systems1
Planning & Managing Change in Data Centres
David CuthbertsonMobile 07717 883177
BSc MBCS MIOD
© Square Mile Systems2
House keeping
• Fire evacuation• Toilets• Refreshments• Mobile phones• Presentation material
© Square Mile Systems3
Agenda9:15 Introductions9:30 Data Centre Management
David Cuthbertson, Square Mile10:00 Bringing Service Awareness to the Data Centre
Paul Dixon, LloydsTSB10:30 High Density Computing
Bob Wilson, Rittal
11:00 Coffee & Tea
11:25 Optimising Data Centre Power Andrew Gibson, Raritan
11:55 Planning and Managing Change David Cuthbertson, Square Mile
1:00 Buffet lunch
© Square Mile Systems4
Square Mile Systems Overview
Fixed Infrastructure(Cabling, Power, Cabinets, Rooms, Buildings)
Hardware InfrastructurePCs, Network, Servers, UPS, Storage, Other
Virtual InfrastructurePCs, Network, Servers, Storage, DBMS
ApplicationsPC, server, mainframe, SOA
ServicesEnd user, infrastructure, supplier
Business ProcessesDepartmental, Company
What you have
What it does
Who it affects
How it is connected
Where it is
Who is responsible
When did it change
How much we have left
What are our costs
Are there single points of failure
UK based – Cirencester, Glos, UKSister company – AssetGen Ltd
Focus on applying asset & configuration management techniques to large infrastructures & data centres
Develop toolsets for end to end systems and service mapping – The AssetGen Range
Integrate existing CMDB / knowledge sources with other toolsets - All technologies!
Training, design, data capture, process development
Industry Bodies BCS Chairman – ITIL Specialist GroupBCS-DCSG – Data Centre Specialist GroupBICSIITSMF
© Square Mile Systems5
Best Practices Training
• Practical Data Centre Management - Part 1– Managing the facility
• Practical Data Centre Management – Part 2– Managing the external interfaces – Change, ITIL, ISO27001, BS25999
• How to Map Services and Systems– Communicating change / incident impact (ITIL)
All courses are one day and can be held on site if desired
Next public courses PDCM 1 26th JanPDCM2 27th JanHTMSS 28th Jan
© Square Mile Systems6
The AssetGen Toolsets
AssetGen ConnectPhysical Infrastructure
AssetGen SysMapLogical
Infrastructure
VISUALISATION
Business ProcessesDepartmental, Company
ServicesEnd user, infrastructure, supplier
ApplicationsPC or server based
Virtual Infrastructure ComponentsPCs, Network, Servers, Storage, DBMS
Hardware Infrastructure ComponentsPCs, Network, Servers, UPS, Storage, Other
Fixed Infrastructure(Cabling, Power, Cabinets, Rooms, Buildings)
© Square Mile Systems7
Data Centre Management Issues
• Technology– Challenges, opportunities
• Capacity– Space, power, cooling, connectivity, people
• Risk– Change, recovery, security, resilience, support
• Organisational– People, policies, processes, communication, cost
© Square Mile Systems8
Technology Requirements
No. of Servers per cabinet 3-6 30-40Power Disipated per cab. 300-2kW 3kW - 25kWCurrent service to cabinet 16A 32 A or 3 phaseTypes of Equipment Servers Blade Servers
Monitor Power Distribution UnitsKVMs MidSpan Boxes
Power Strips Disk Arrays (Storage)UPS Smart Power Strips
Network types 100Base-T 1G, 10G, SANNo. of Cables Power 1 or 2 2 to 6(per server) Network 1 or 2 5 to 10
Total 20-30 300 - 400
© Square Mile Systems9
New Technology Challenges
Sun Blade 8000 Blade Chassis– 4 Power supplies (N+1) 9kW– 3 chassis per rack
HP C7000 Blade Chassis– Up to 6 Power Supplies 13kW– 4 chassis per rack
Cisco Nexus Data Centre Switch– 3 Power Supplies 12kW
© Square Mile Systems10
Same components, different approach
© Square Mile Systems11
Understanding Position and Orientation
Raised Floor
Cold Air
FrontFront Front Front
Server
Warm Air Server
Hot AirServerMeltdown!
Cold Air Cold Air Cold Air
© Square Mile Systems12
Hot & Cold Aisle TechniqueCeiling
Extractor
Raised Floor
Cold Air
Warm Air
Front Front FrontFront
Cold Air
“HOT”“COLD” “COLD”
© Square Mile Systems13
In Practice – its more difficultCeiling
Extractor
Warm Air
Cold Air Cold Air
Raised Floor
Front Front
“HOT”“COLD” “COLD”
FrontFront
Recirculatingcurrents
Power TrayIncoming Cold Air Cable Tray
OverheadPower Tray
Recirculatingcurrents
Reduced airflow
© Square Mile Systems14
Its not just a “room”
X
“Hot” “Cold” “Hot”
© Square Mile Systems15
Power View
Loading
>6kW
2-6kW
<2kW
© Square Mile Systems16
Different Power Views
LINK 10/100FEATURE
LANSERIAL
CURRENT���������������
ON = I OFF = U
BLINK = REMOTE
OUTLET #I/U TOGGLE
RESERVED
STATUS 9 10 11 12 13 14 15 16
1 2 3 4 5 6 7 8
100-240V
~
50~60Hz
1.2A
KVM
Servers
LINK 10/100FEATURE
LANSERIAL
CURRENT���������������
ON = I OFF = U
BLINK = REMOTE
OUTLET #I/U TOGGLE
© Square Mile Systems17
Increasing System ComplexityDEALVIEW INTERNAL AUDITS
GLOBAL INS FUNDING PAYMENTS CASH
MANAGER CHECKBAL INCOMINGFUNDS
SECURE GH
SECURE EDI INVEST
Issue: Complex systems arenot easy to comprehend or validate
© Square Mile Systems18
Common Sense
To manage change:
1. Establish a baseline2. Manage change using process3. Verify processes are working
Planning installs and moves is more complexOptimising power and cooling is now requiredPredicting change impact is taking longer due to increased risksManaging a data centre requires a proactive management style
Or… Disruption, delays, increased operating costs
© Square Mile Systems19
Improving Controls
• Environment limits• Information sets - formal and informal • Working practices - formal and informal• Roles / responsibilities• Current issues• Establish priorities
© Square Mile Systems20
The EU Code of Conduct on Data Centres
Proposed by the BCS-DCSG (Data Centre Specialist Group) to avoid unnecessary legislation. Submitted 19th November 2008. http://dcsg.bcs.org
Covers A. IT energy loadB. Facilities energy load
Aims Metrics to measure current efficiencySet energy efficiency targetsSupport procurement of energy saving componentsRecommend working practices
Obligations Data collected monthly, reported annually to EU
© Square Mile Systems21
EU Code of Conduct – Best Practices
All practices categorised asA. OptionalB. ExpectedC. Expected for new installs or replacementD. Expected for new data centre build or significant refurbishment
Covers many issues – Cooling, power, devices, management, procurement, data storage
© Square Mile Systems22
Summary
Data centres are becoming more complex, densities of cabling and equipment are increasing and the business impact of a failed change is greater. Plus the “green” issue
Changes in data centre management• Data centre teams (and managers) need additional skills• Complexity needs managing – capacity, connectivity, services• Working practices must evolve• Evidence of control over critical systems will be demanded
- internal teams, suppliers, outsource partners
© Square Mile Systems23
© Square Mile Systems24
© Square Mile Systems25
Planning & Managing Change
© Square Mile Systems26
Common Data Centre Challenges
1. Quickly determining the impact of infrastructure changes (planned or reactive)2. Speed of provisioning3. Knowing if we will exceed design limits for power, cooling or recovery capability4. Maintenance planning needs – UPS, PAT, load balancing, decommission5. Disparate data sources which overlap, disagree or have gaps
- Excel, BMS, Data Centre management tools, monitoring tools, inventory
6. Capacity management and supporting controls7. Optimising existing space and capacity 8. Interfaces with deployment, platform, service, partners, teams9. Skills of existing teams – technical, process10. Lack of integration with Configuration Management initiatives11. Articulation of management issues12. No baseline to manage from
© Square Mile Systems27
Getting ComplexLINK 10/100
FEATURE
LAN SERIAL
CURRENT� � � � � � � � � � � � � � �
ON = I OFF = UBLIN K = REMOTE
OUTL ET #
I /U TOGGLE
RESERVED
STATUS 9 10 11 12 13 14 15 16
1 2 3 4 5 6 7 8
100-240V
~
50~60Hz
1.2A
KVM
Server
Firewall
Switch
Storage
LINK 10/100
FEATURE
LAN SERIAL
CURRENT� � � � � � � � � � � � � � �
ON = I OFF = UBLIN K = REMOTE
OUTL ET #
I /U TOGGLE
CopperFibrePower
Managing change at equipment level also requires management of connectivity!
Or monitoring results can’t be interpreted
© Square Mile Systems28
What Do We Need To Know?
• Device Inventory– Anything that takes up space, power or is connected through cabling– Includes daughter cards, blades, KVMs, power strips, patch panels etc.
• Space– Amount of space taken up by components– Position and orientation within computer room
• Connectivity– Power– Network– SAN– Other
• Coordination with other data sources– Service, ownership, monitoring etc.
© Square Mile Systems29
Practical Issues
• Who owns the problem of creating and maintaining an end to end data centre knowledge base?– Facilities?– IT Data Centre teams?– Platform teams?– Service management?– Development teams?
• Where do you start?– People– Process – Toolsets
Is this going to be solved by ITIL configuration management and a CMDB?
© Square Mile Systems30
Different Teams, Different Focus
Business ProcessesDepartmental, Company
ServiceManagement
DataCentre
NetworksLAN/SAN
Applications
Mid-range Servers
Systems
DesktopsIMAC
ServicesEnd user, infrastructure, supplier
ApplicationsPC, server, mainframe, SOA
Virtual InfrastructurePCs, Network, Servers, Storage, DBMS
Hardware InfrastructurePCs, Network, Servers, UPS, Storage, Other
Fixed Infrastructure(Cabling, Power, Cabinets, Rooms, Buildings)
© Square Mile Systems31
Network Connectivity VariationsServer Switch
PP/FB PP
PP/FB PP PP
PP PPPP PPPP/FB
Server Cabinet
Switch Cabinet
1. Point to Point
2. Intercabinet link
3. Cabinet to wiring zone
4. Between rooms, areasVertical Wiring
Patch cable
Storage Terminated in patch panel
© Square Mile Systems32
In Reality
• Many informal practices and walking databases• Lack of clarity around ownership and interfaces between
projects/operations teams• Complexity of infrastructure documented in many disparate
formats– Power, networks, storage, space, servers, etc.– Data sets centred on teams– End to end view difficult to get
• Funds for projects, not for management
© Square Mile Systems33
Moving Forward
Before After
Excel
Word
Visio
ExcelExcel
Visio
Visio
Word
Word
Word
WordVisio
Visio
Visio
Excel
AssetGenSystem
DifferentViews
Reduce effortShare informationRe-use dataStandardisationTrustedProject/Operations useHome/central working
Visio
Excel
Word
© Square Mile Systems34
The AssetGen Range
AssetGen ConnectPhysical Infrastructure
AssetGen SysMapLogical
Infrastructure
Business ProcessesDepartmental, Company
ServicesEnd user, infrastructure, supplier
ApplicationsPC or server based
Virtual Infrastructure ComponentsPCs, Network, Servers, Storage, DBMS
Hardware Infrastructure ComponentsPCs, Network, Servers, UPS, Storage, Other
Fixed Infrastructure(Cabling, Power, Cabinets, Rooms, Buildings)
© Square Mile Systems35
AssetGen In More Detail
SysMapExpert
Service MappingVisualisation- visio, netViz
SearchingImpact formsPath tracingRack diagramsCapacity - space, power, portsInventoryWorkflow Audit trails
WebBrowser
AssetGen ConnectPhysical Infrastructure
AssetGen SysMapLogical
Infrastructure
MS SQL Database
Service DeskCMDB
(if suitable)
refresh
MonitoringDiscovery
AssetSpreadsheets
refresh
PC
Planner Space planningVisualisation - visio, netViz
Impact Analysis
© Square Mile Systems36
AssetGen Planner
AssetGen ConnectPhysical Infrastructure
AssetGen SysMapLogical
Infrastructure
MS SQL Database
AssetGen Planner
Data Center Planning
1. Impact Analysis
2. Space Planning
3. Visio Diagrams
4. Capacity Reporting
Location and rack changeLAN/SAN//Power connectivityMultiple services and software
Finding equipment spaceLAN/SAN//Power resourcesDesign criteria
Top down floor plansRack layoutsEmbedded device data / hyperlinks
Space and Power capacityLAN/SAN//Power resources
© Square Mile Systems37
Change ImpactBusiness Processes
Departmental, Company
ServicesEnd user, infrastructure, supplier
ApplicationsPC, server, mainframe, SOA Stage 2
Service ImpactVirtual Infrastructure
PCs, Network, Servers, Storage, DBMS
Hardware InfrastructurePCs, Network, Servers, UPS, Storage, Other
Stage 1Hardware
ImpactFixed Infrastructure
(Cabling, Power, Cabinets, Rooms, Buildings)
© Square Mile Systems38
Change Impact Analysis - Example
1.Start Point
2. Initial Inventory
3. Trace Connectivity
4. List Target Devices
5. Hardware Impact
Stage 1Identify Hardware Impact
Endpoint Tracing
7. Select Impact Criteria
8. Select Dependency
9. Service Impact
6. Manual De-Select
10. Import List (Excel)
Stage 2Identify Service Impact
(Optional)
© Square Mile Systems39
PDU
Hosts Power Strips
Circuit Breakers
IntegratedRacks
Active Equipment
Hosts
Hosts Hosts
SAN LAN
PowerDependency
Direct Connect
PDU or Room Power Down Example
NetworkDependency
© Square Mile Systems40
Infrastructure Upgrade Example
Hosts
Hosts
Hosts
Hosts
Hosts
Core Switches toBe upgraded
Edge Switches
© Square Mile Systems41
Change Freeze Example
Host B
Power Strips
Critical Hosts
Host A Host C
SAN LAN
Host D Host E
BackupBladeChassis
Supporting Infrastructure
Identify the supporting infrastructure for a critical service!
© Square Mile Systems42
Finding Suitable Space
How to quickly find and reserve suitable cabinet space- Project needs or specific equipment installs
May involve a combination of design criteria1. Room and cabinet selection2. Available local connectivity3. Power consumption4. Standardised design criteria
© Square Mile Systems43
Generating Visio Diagrams
Quick and easy distribution of rack and device details
1. Room ViewsTop down views for overlay onto grids, cable runs etc.Embedding of cabinet dataHyperlinks into AssetGen Connect
2. Cabinet ViewsRack views with colour coding of work order statusSelect individual, groups, rooms, buildingsEmbedding of device data Hyperlinks into AssetGen Connect and AssetGen SysMap
© Square Mile Systems44
Room View – Example
01-01
1095
01-02
505
01-03
2405
01-04
5655
01-05
2405
01-06
2055
01-07
7455
01-08
1705
01-09
1455
01-10
5
02-01
1095
02-02
505
02-03
255
02-04
255
02-05
3505
02-06
255
02-07
1555
02-08
2255
02-09
2755
02-10
5
Each Cabinet has hyperlink and data embedded from AssetGen
Icon shows if Cabinet is within power budget
Double- Click on cabinet to go to rack
view
© Square Mile Systems45
Cabinet View ExampleFront
PP01-03-01
SVR-BHAM-010301
UK_BIRM_UX01
SERVERWIN0001
Rear
Birmingham | Call Centre - UK | Computer Room | 01-03
Power strips and devices positioned
automatically
Device data embedded
automatically
Colour coding shows work
order in progress
Hyperlinks back to Connect or
SysMap
© Square Mile Systems46
AssetGen Planner – Topology Views
Automated generation of topology views to quickly show resilience, single points of failure and paths
• Technology or platform views– Power infrastructure– LAN/SAN/WAN Network
• User selection of location, equipment and connectivity– Multi-technology views– Supporting infrastructure for critical systems
• Output in Visio
© Square Mile Systems47
Return on Investment
• Direct Cost Benefits– Unused inventory– Audits– Site surveys– Optimising resources – space, power, cooling, connectivity– Less reactive troubleshooting – Team workload for coordinating and planning changes– Capacity management and reporting
• Indirect Cost Benefits– Cost of disruption on business users and customers– Confidence in controls and reporting
© Square Mile Systems48
Typical Approach to Baseline a Data Centre
• Fixed Infrastructure– Power, Cabling, Cabinets, Rooms etc.
• Active Components– Servers, Networks, Storage, Firewalls, Software, Services
• Map dependencies between all– Patching, power, LAN, SAN, WAN, data flows, process steps
© Square Mile Systems49
Planning and Managing Change Summary
• Changing requirements for Data Centre management• Seen the latest technology that can cover
– Space– Power– Connectivity– Workflow– Links into service management data
• Understand where Square Mile can help– People Skills and communication– Process Techniques and practices– Toolsets AssetGen and integration