Page 1
White Paper – Dynamic Rack Control Dipl.-Kffr. (FH) Csilla Burján Dipl.-Ing. Markus Schmidt
Copyright © 2009 All rights reserved.
Rittal GmbH & Co. KG
Auf dem Stützelberg D-35745 Herborn
Phone +49 (0)2772 505-0
Fax +49 (0)2772 505-2319 www.rittal.de
www.rimatrix5.de
Page 2
Dynamic Rack Control
Contents
1. Introduction.......................................................................................................................5
2. RFID .................................................................................................................................5
2.1. Technology ...............................................................................................................5
2.2. Strengths and opportunities......................................................................................6
2.3. Acceptance...............................................................................................................7
3. Optimisation potential at the data centre ..........................................................................8
3.1. Stocktaking ...............................................................................................................8
3.2. Power supply and cooling.........................................................................................8
4. Description of the Rittal Dynamic Rack Control system ...................................................9
5. Benefit aspects...............................................................................................................12
6. Prospects .......................................................................................................................13
2
Page 3
Dynamic Rack Control
List of illustrations
Fig 1 – RFID global market sales development in USD mn .....................................................7 Fig 2 – RFID significance over the next few years ...................................................................8 Fig 3 – Aerial with lit LEDs .......................................................................................................9 Fig 4 – Rittal tags and aerial.....................................................................................................9 Fig 5 – Basic data and power balance of an individual rack extension ..................................10 Fig 6 – Master record of a 482.6 mm (19”) component..........................................................10 Fig 7 – Supplementary management information ..................................................................10 Fig 8 – Technical layout of DRC.............................................................................................11 Fig 9 – PSM rail with LEDs per socket ...................................................................................12 Fig 10 – Costs incurred by downtimes ...................................................................................13 List of tables Table 1 – Active and passive transponders compared ............................................................6 Table 2 – Properties and advantages of RFID .........................................................................6 Table 3 – Opportunities and advantages of RFID technology..................................................7
3
Page 4
Dynamic Rack Control
List of abbreviations
CMC - Computer Multi Control
CPU - Central Processing Unit
DRC - Dynamic Rack Control
HU - Height Unit
HF-RFID - High Frequency Radiofrequency Identification
IT - Information Technology
RFID - Radiofrequency Identification
DC - Data Centre
UPS - Uninterruptible Power Supply
Tagging - Data Identification
4
Page 5
Dynamic Rack Control
1. Introduction
The internet of things – not only computers, also a growing number of objects and processes
are being networked. The basis for this is provided by RFID technology. Radiofrequency
identification or RFID opens up new prospects for a great many sectors and is therefore
evolving into a cross sectional technology. A pioneering role is adopted above all by logistics
and inventory control, with great potential and possible applications still to spare.
One of these many sectors involves the optimisation of data centres. To date little
consideration was given to physical stocktaking with RFID in conjunction with monitored
power and cooling requirements. In this respect the further development of RFID technology
and its innovative application can contribute towards generating solutions for a wide range of
diverse requirements.
2. RFID
2.1. Technology
RFID makes possible the contactless identification of objects and persons by means of radio
waves. An RFID system infrastructure consists of a transponder, a transceiver, and an IT
system operating in the background. At centre stage is the transponder, a microchip with an
aerial coil. This is integrated in a carrier, e.g. a plastic card. As a rule a numeric code is
stored in the chip. This encrypts information that is stored in a database. Consequently every
object with an RFID transponder receives an unmistakable identity.1
The information is read out and written by means of radio waves. At low frequencies this is
an inductive process via a near field, at higher frequencies via an electromagnetic far field.
The distance over which an RFID transponder can be read out depends on the design
(active/passive), the frequency band used, the transmitting power, and environmental
factors.2
Search procedures applied to these objects are similar to those for information on the
internet. The tag transmits data, and the reader forwards the code to a database. The IT
system decrypts the code and links this to information stored in the database or on the
internet. The knowledge or the intelligence of the system does not lie with the transponder,
but in the databases.3
Today there are transponders in the most diverse shapes and sizes. Depending on the
application active or passive transponders are used.
1 http://www.info-rfid.de/downloads/basiswissen_rfid.pdf 2 http://www.info-rfid.de/downloads/rfid_leitfaden.pdf 3 http://www.info-rfid.de/downloads/basiswissen_rfid.pdf
5
Page 6
Dynamic Rack Control
Active Passive
Power supply battery radio waves
Service life dependent on battery unlimited
Price high low
Storage space large small to medium
Writability rewritable once only or rewritable
Range far few cm to several m
Read rate medium to high low to medium
Table 1 – Active and passive transponders compared
2.2. Strengths and opportunities
Owing to the great many advantages industry and trade are today backing RFID technology.
Table 2 presents the properties of RFID data media and their practical advantages.
Properties of RFID data media Practical advantages Every chip has a unique serial number that
is assigned once only in the world (UID).
Unique assignability on single product level
Complete product individualisation
Rewritable data storage in the chip Flexible data management directly on the
product
Information on the RFID data media can be
modified, deleted, and supplemented at any
time.
Product, maintenance, production, and service
data available directly on the product
Contactless communication out of the line
of sight between RFID data medium and
read/write system
Insensitivity to soiling owing to chip integrated
at protected sites
Invisible integration in a great many existing
products
Simplified process automation
Virtually 100% first read rate Accelerated processes
Reduced misreads
Simultaneous reading of more than one
RFID data medium in the one operation
(multiple capture)
Accelerated processes
Reduced misreads
Table 2 – Properties and advantages of RFID4
4 http://www.smart-tec.com/index.php?id=2273
6
Page 7
Dynamic Rack Control
RFID technology presents opportunities for all public sectors – for the economy, science,
public institutions, and recreation. Some examples are listed in the following Table 3.
Opportunities Example
Optimised processes Goods and pallet identification in logistics
Simplified traceability Organic foodstuffs
Assured authenticity Electronic passport, animal identification
Improved product safety Drugs
Optimised inventory management Automated stock replenishment control at supermarkets
Simplified access control procedures Chip cards
Table 3 – Opportunities and advantages of RFID technology
2.3. Acceptance
In spite of a number of setbacks RFID has turned out to be a success on several sectors.
This finds testimony in a great many projects successfully applying RFID, for instance goods
tagging and transponder assisted theft protection. Also the market figures are promising.
Basing its calculations on the 2007 global turnover of USD 917mn the consulting company
Gartner expects strong international growth over the next few years. They predict that growth
in 2012 will reach USD 3.5bn.5 Fig 1 presents the expected trend of global turnover.
Fig 1 – RFID global market sales development in USD mn
How companies accept and rate the technology was investigated in 2008 by the FTK
research institute for telecommunications in collaboration with the RFID information forum.
The findings – two thirds of the 298 investigated companies, most of them SMEs, are well or
very well informed of the essential aspects of RFID. The majority of the investigated
companies agree that in the next few years RFID will gain greatly or very greatly in
5 http://www.networkcomputing.de/gartner-erwartet-12-milliarden-dollar-umsatz-mit-rfid/
7
Page 8
Dynamic Rack Control
significance for the Mittelstand (SME), the sector, and their own company. Fig 2 presents the
surveyed ratings in the form of a bar chart.
Fig 2 – RFID significance over the next few years
In the face of these evaluations we must now start closing the information gaps and
presenting new, potential applications. The intention behind Dynamic Rack Control is to tap
improvement potential at the data centre and present new approaches to finding solutions.
3. Optimisation potential at the data centre
3.1. Stocktaking
At data centres – in particular above a certain size – it is difficult to maintain an overview of
the installed hardware components. Although it is possible today to communicate with every
intelligent IT device, problems are encountered with the physical assignment between the
rack and the affected height unit. Also the device structure in each enclosure with servers,
fans, UPSs, etc., often lacks transparency. Against this background the processes involved
in taking stock and constantly updating data on the distribution of components at the data
centre are complex and in most cases time consuming to boot. In many cases the available
documentation generated by hand is not examined for correct contents. These details though
are necessary if the right decisions are to be made in the event of a fault.
Another problem is the half life of the captured information. Capture and updates always
depict an instantaneous record of the DC inventory. Efficient rack assignments and
transparent component administration though require up to date and therefore reliable data.
3.2. Power supply and cooling
Fitting racks with components is all too often based on experience, with mostly no
consideration to an efficient power supply and cooling or adaptation to changing power
requirements. This gives rise to hot spots and the probability of faults becomes greater.
When for instance a server with key applications signals a fault, availability is safeguarded
when virtual machines are relocated to other racks. Yet which rack houses the server, and
8
Page 9
Dynamic Rack Control
which provide the optimal physical conditions (electrical power / cooling)? How much
electrical power is drawn by a device in the enclosure, and what are the present cooling
requirements? Answers to these questions require up to date and reliable information on the
device structure in each rack and their physical conditions. The following presents the
Dynamic Rack Control system and illustrates its benefits and advantages for the
demonstrated optimisation potential.
4. Description of the Rittal Dynamic Rack Control system
Dynamic Rack Control (DRC) utilises RFID technology directly in the rack for the contactless
and realtime capture of all components fitted in the 482.6 mm (19”) level. The concept –
components fitted with RFID tags constantly transmit information to a strip aerial installed in
the enclosure. This features a sensor at every mounting hole so data can be retrieved from
every height unit. In this manner every single server is captured free of contact immediately
when it is installed. In addition all basic information (hardware maker, component type,
installation date, software) can be stored on the RFID tags for permanent access. Figs 3 and
4 depict the illuminated strip aerial and an installed tag respectively.
Fig 3 – Aerial with lit LEDs
Fig 4 – Rittal tags and aerial
Flashing LEDs at the sensor confirm when the tag is identified, read out, or written to. In this
manner an up to date overview can be gained of all components installed in the rack, sorted
according to device class, power consumption, or form factor.
The rack configuration can be viewed on the one hand as a visual representation on the
CMC TC website6, on the other as a list of data allowing import and export via the XML file
format. These records can then be processed further e.g. in external databases and simplify
to a great extent the daily work involved in optimising a data centre.
6 http://www.rittal.de/downloads/PrintMedia/PM5/de/cmc_tc_wireless.pdf
9
Page 10
Dynamic Rack Control
Fig 5 – Basic data and power balance of an individual rack extension
Fig 6 – Master record of a 482.6 mm (19”) component
Fig 7 – Supplementary management information
Rittal applies HF RFID technology (f = 13.56 MHz) whose carrier frequency allows
particularly compact sizes for the aerial modules. The data are written to the tag in
compliance with ISO 15693. These can be captured and edited with suitable handheld
readers, both when the tag is integrated directly in the rack and when the components have
been removed.
At a later time an optional RFID reader in the enclosure frame and the corresponding tags at
the base will also make it possible to capture the exact rack position at the data centre. It is
therefore recommended, above all with rack extensions and new installations, to take this
innovative inventory management system directly into consideration. Fig 8 depicts the
technical layout of DRC.
10
Page 11
Dynamic Rack Control
Fig 8 – Technical layout of DRC
The option of storing a great many component characteristics on their passive RFID tags
helps to identify free capacity on the 19” level in racks and to draw conclusions on the
installed power consumers. Every rack therefore can be fitted with additional hardware for
the optimal space utilisation, cooling performance, and power supply. These data are
available at all times up to date and online, and modifications are applied immediately: there
is none of the manual capture or stocktaking otherwise usual.
RFID is the basis of a module for an extensive monitoring system. This can also be
integrated in RiZone, or the Rittal Infrastructure Management Suite. RiZone is a
management platform that has been optimised to the physical infrastructure of a DC. It
manages inhouse infrastructural devices like CMC; UPSs; power supplies in the rack, room,
and floor power outlets; cooling up to the rack level; etc.; but can also communicate with
other products via SNMP and BACnet.
This makes possible
the administration of the entire physical infrastructure of a data centre
measurements of the power draw and individual servers switched on and off per slot
statistics
characteristics for power draws, etc.
11
Page 12
Dynamic Rack Control
Fig 9 – PSM rail with LEDs per socket
Dynamic Rack Control, RiZone, and management software like e.g. Microsoft's System
Center Operations Manager, can be used to observe and control all key parameters like
cooling, power supply, and energy consumption at a data centre.
In addition to enhanced failure safety and energy efficiency, applications can be relocated
faster, devices found and replaced more quickly, and optimisation potential utilised better in
the event of a fault.
5. Benefit aspects
With RFID technology Dynamic Rack Control offers a great many benefits at data centres.
Components must be located, the device type and serial number registered, the records
sorted, and the metadata (purchase date, maintenance, etc.) updated at regular intervals.
Depending on the size of the DC and the available staff these tasks are a considerable time
and coordination factor – not only at own data centres, but also for outsourced processes.
DRC offers here the advantages of automated stocktaking. It automatically identifies
components and assigns these to the rack and exact height unit. Changes to hardware
positions are registered and signalled automatically.
Furthermore human error and their effects on hardware stocktaking play a role. Incorrect
identification of components' installed locations and inadequate documentation may have far
reaching consequences in the event of failure: decisions, e.g. to shut down a server, are
made on the basis of wrong information and can cause considerable financial losses. The
costs incurred by the non-availability or downtimes of a service depend heavily on the
affected application and business segment. Fig 10 depicts the estimated costs per hour in
US dollars on various sectors.
12
Page 13
Dynamic Rack Control
28 90 90 113
1100
2500 2600
6500
0
1000
2000
3000
4000
5000
6000
7000
Manufa
cture
Logis
tics
Retail
Home s
hopp
ing
Media
(pay p
er vie
w)
Bank d
ata ce
ntre
Credit c
ard pr
oces
sing
Brokeri
ng
Cos
ts /
h (in
TU
SD)
Fig 10 – Costs incurred by downtimes7
Nor must we forget the monetary effects on IT service providers and their service level
agreements (SLAs). SLAs are agreements between customers and service providers for
requested and provided services like availability, response times, online times, etc. In this
context non-availability of services can affect a great many customers and processes.
The loss of customers owing to loss of image, compensation for inadequately provided
services e.g. in the form of costs for additional purchases, possible liability for consequential
damage to partners and suppliers, etc., are virtually unquantifiable effects that can also
occur.
By providing hardware data and integrating them in existing monitoring systems Dynamic
Rack Control can simplify asset management at the DC. Other possibilities like the
management of electrical power draws and losses for every component, conclusions on free
capacities for cooling and power supply requirements in the rack, and as a further step the
integration in RiZone are provided by an extensive physical DC management tool.
6. Prospects
RFID links the physical world with the virtual world of digital data. Although the internet of
things is still very much in the future, RFID already plays a significant role in our times. Many
RFID projects have already become reality, e.g. for capturing vehicles for toll systems in
7 Contingency Research,
http://www.tecchannel.de/server/hardware/458076/it_systeme_ausfallsicherheit_im_kostenvergleich/index3.html
13
Page 14
Dynamic Rack Control
14
Dubai, or for registering epasses, etc. According to Gartner the value of this technology lies
not only in RFID, but in the innovation of business processes.
There appears to be a demand for this. According to a 2008 study8 for the economical use of
RFID 41 of 149 subjects surveyed could well or very well imagine applications in the field of
inventory management. On the other hand those surveyed could imagine RFID less well for
status checks. This shows that the advantages of this technology are still not known to an
adequate extent. What we must now do is to make companies aware of the benefits of RFID
at data centres and to demonstrate the advantages.
8 International Performance Research Institute, "Economic use of RFID", 2008