isx datacenter chall
TRANSCRIPT
APC InfraStruXureTM
Thomas LamSenior District Manager, APC HK(852) [email protected]
Agenda• DataCenter Challenges • Introduction to InfraStruXure
– Rack - Cooling– Power - Services
• ISX Reference Sites• When do you need ISX?• Open Discussion
Environment
HVAC Power Safety/Security Structure Building Management System Data Center Infrastructure
Information Technology
Network Server Processes Operating Systems Data Management Security Applications
People
Human Resources Management Operations Recruiting Training
Process
Tools Organization Metric Knowledge Base Continuing Education
NCPI – 4 Availability Layers
Server Room/ Datacentre Challenges
Data Center ChallengesBackground Information
Voice of Customers500 Client Visits
Engineering User GroupsIT Managers/consultantsConsulting EngineersFacility ManagersProject Managers Electrical Contractors
Professional 3rd Party Consultants
Availability Consulting & Services
What are the top three things you would change about your data
center infrastructure if you could?
Data Center Basic Design
Top Data Center Challenges
Top Data Center Challenges1. Optimize capital investment and available
space“We made 100% investment on Infrastructure
although we needed only 20% infrastructure at the beginning.”
• Lack of scalable solutions• Cannot “right size”• High maintenance / service costs • Infrastructure takes too much space
Top Data Center Challenges 2. Adapt to ever-changing requirements
“I consolidated five servers into one more powerful high-availability server. The new server has 2 / 3 power supplies and my infrastructure cannot support that ”
• Small changes require too much planning• Today’s infrastructure is difficult to re-locate• Increasing power density causing heat issues (42
servers in one rack!)• New Servers do not fit into old Racks
Top Data Center Challenges 3. Plan for a power density that is
increasing and unpredictable. “Servers are continuing to get smaller. I continue to bring in more power and cooling as my IT equipment changes.”
• Future power densities cannot be predicted
• Power infrastructure should be as flexible as IT systems
• Manage heat to avoid downtime
Top Data Center Challenges 4. Accelerate speed of deployment
“I need a new computer room in two weeks and it takes me six months to get it built.”
• Customization leads to long design time• Multiple vendor interfaces make
planning difficult• High level consulting required• Installation needs long planning and is
labor intensive
Top Data Center Challenges5. Lower the cost of service contracts
“I bought the 3-phase UPS system at a great price, but I was locked into an expensive service contract.”
• Servicing a UPS should be as easy as IT equipment
• Preventative failure analysis• Service costs should be “right sized” and
match the current infrastructure being used.
Top Data Center Challenges 6. Increase availability per dollar
“If I want redundancy in my power infrastructure it takes twice the space and cost.”
• Infrastructure Availability lower than IT systems Availability
• Redundancy too expensive (dollars and space)• Long runtime expensive for whole data center• “Spend or sacrifice” are today’s choices• Need “targeted availability”
Top Data Center Challenges 7. Decrease Mean-Time-To-Recovery
(MTTR) “Failure of my data center infrastructure means we will never achieve our availability goals.”
• Build in redundancy like RAID• Make systems smart
– Self-diagnosing– Predictive failure analysis
• Make systems easy to service
Design Power Capacity and requirement over the lifetime of a datacenter
0%
20%
40%
60%
80%
100%
120%
140%
160%
0 5 10
Years From Start-up
% o
f C
apac
ity
Nameplate Power Capacity
Installed Power Capacity
Expected Power Requirement
Actual Power Requirement
70% Excess Costs - Wastage
Top Data Center Challenges 8. Minimize points of failure between the UPS
and critical load
“Even though my computer room is supported by a centralized UPS, I still see downtime when problems occur in the distribution panel feeding my racks.”
• Reducing the number of components between the UPS and the critical load will increase availability
Integrated Power, Cooling, and Rack Infrastructure for
Critical Networks
Increase availability while decreasing your total cost of ownership
Rack
Have you ever seen a site like this...?
What is APC’s Rack System?
Power (PDU/UPS)Cooling
Rack Accessories
Management
Cable Management
NetShelter VX Core Advantages
• Cooling• Power Distribution• Cable Management• Management• Guaranteed Compatibility
Compatibility“Fits like a Glove” Guarantee
• All EIA-310 19” equipment will fit in NetShelter VX or your money back.
• APC is committed to ensuring that all RM equipment physically fits into the NetShelter VX.
Integrated Cable Management
Power
Centralized (Room)- one UPS and/or central or distinct
location of the UPS
Power Protection Methodologies in the Data Center
Point-of-Use (Rack)- one UPS used to protect a single
device or rack
Zone (Row)
- a UPS (along with other components) powering a row(s) of racks or a group of non-racked equipment
Power Protection Methodologies in the Data Center
Type C
Type B
Type A
Large to Very Large Data
Centers
3Ø Power
>200kW
>500m2
Medium to Large Data Centers
1Ø or 3Ø Input Power
12-200kW
50-500m2
Small Data Centers
1Ø Input Power
2-12kW
10-50m2
…APC has 3 standard solutions types depending on the size of the Zone:
Scalable Data Center Architecture for Infrastructure On-DemandTM
• 40kVA/40kW capacity in 10kVA/10kW power module increments
- power factor corrected input
- fully rated inverter
• Power modules arranged in N+1 configuration
• Battery modules arranged in a parallel redundant configuration
• All modules are hot-swappable
• Housed in a 870mm NetShelter VX enclosure
- H x L x D = 2070mm x 610mm x 870mm
• 125% continuous rated, static switch module
• XR frames available for longer runtime requirements
Symmetra PX 10 - 40 kW
Scalable Data Center Architecture for Infrastructure On DemandTM
Replaces
■ Isolation Transformer
■ Service Bypass Panel
■ Output Panel, Cabling, Wiring
■ Standard 19” cabinet footprint
Power Distribution System
Availability Tiers:
Choose your level
N+1
N+1 with External Redundancy
Arrayed Redundant
On-demand Scalable, Manageable, Pre-engineered Solutions
Tier I - N+1 Design Statistical Availability = 99.99%
Tier II - External Redundant Design
Statistical Availability = 99.999%
Tier III - Arrayed Redundant Design Statistical Availability = 99.99999%
Cooling
Rack Air Distribution Unit (ADU)
• 2U rack-mounted fan unit that delivers cool air to the equipment contained in a rack enclosure.
• ADU helps to eliminate temperature differences between the top and bottom of the enclosure.
• Prevents hot exhaust air from re-circulating to the inlet of the enclosure.
• Best suited for racks with load densities up to 3kW.
ARU
Air Removal: ARU
NetworkAIR Air Removal Unit (ARU)
Applications:
• Densely-populated racks in small data rooms to large data centers
• IT enclosures with cabling that blocks air exhaust in the rear
• Enclosures that exceed 3.5 kW load capacity
• 8.0 kW heat removal capacity at 1200 cfm (2040 m³/h) design airflow with one fan as redundancy and 11°C temperature rise across servers
• Drop ceiling environments where the A/C efficiencies can benefit from higher return temperatures
NetworkAIR FM
• Floor mounted precision air conditioning for IT and data center environments in 35, 40 and 50 kW modules. (10, 12, 15 tons)
• The next generation in precision cooling systems.– More efficient– More reliable– Less expensive to maintain– More intelligent
Services
ISX Reference - HK
• Disney Hong Kong
• NRI Limted
• Poly University
• City University
• Education Dept.
• Agile Software Ltd.
• Radio HK
• Kerry Logistics
• Macau Customs
• Department of Health
• Gold Quest
• Hospital Authority (EMSD)
Thank You!
Open Discussion !