Industrial Ethernet
Infrastructure Design Seminar
TP02:Implementing
Network Redundancy
Standardised vs Proprietary Technology
• Standardized:
• Works across manufacturers
• Future-proof
• Well understood
• Proprietary:
• Fulfils niche requirement
• Better performance (mostly)
• More simple
A few incompatible ring technologies
X-Ring
N-Ring OnTime-Ring
ICON
M-Ring P-Ring
HSR
RapidRing
Rapid Super
HIPER-Ring Ring
Real-Time Ring
FRNT
S-Ring
Take Note…..
• There is no standard for measuring network
recovery time
• Exact meaning of “recovery”?
• Network load?
• Number of Learned Addresses?
• Location of failure?
• Source to destination, or round trip?
• Type of traffic?
• Interaction with other (redundancy) protocols?
• …
Two Stages of Redundancy
• Whichever redundancy method is used, there
are two stages
• Re-establish the physical connection
• Re-establish the logical connection
Re-establishing Physical Communication
A
PC1
B E
C D
PC2
Re-establishing Logical Communication
A
PC1
Learned Address
Table
E Switch “D”
B
Device Path
PC1 Port 1
PC1 Port 2
C D
Port 1
PC2
End device redundant connections
• The redundancy functionality must be provided by
the end devices
• This is not a function of the network equipment
Ethernet Ethernet
Network Network
Rapid Spanning Tree Protocol
Objective
• Creation of resilient meshed networks
HIRSCHMANN HIRSCHMANN HIRSCHMANN HIRSCHMANN HIRSCHMANN HIRSCHMANN HIRSCHMANN
HIRSCHMANN
Graph Theory
Spanning Tree Graph
Rapid Spanning Tree Protocol
Advantages
• Standardized (802.1D-2004)
– Works across all manufacturers
• Supports multiple failures
• Simple to implement
– In theory automatic configuration, but usually needs to be fine-tuned
• Supports loop prevention
Disadvantages
• Unpredictable recovery times
– From milliseconds to seconds
• Unsuitable for large rings
– Maximum 40 hops
RSTP Q & A (1)
Question:
• When should I recommend RSTP?
Answers:
• If the application must tolerate multiple network failures
• If the customer wants a standards-based solution (but see MRP, HSR, PRP)
• Other:
– Fast but not consistent recovery times are required
– Small network diameter
– Connection to an existing RSTP network (not exceeding RSTP specifications)
RSTP Q & A (2)
Question:
• Can I use some unmanaged switches in my RSTP
network?
Answers:
• Yes – on the edge of the network, with no redundancy
• Yes – in the core of the network, but recovery times will be
unpredictable
Overall – if you do this, your RSTP network will be outside the
specification. It is not recommended.
HIPER Ring
Objective
• Creation of a resilient ring structure
HIRSCHMANN HIRSCHMANN HIRSCHMANN
HIRSCHMANN
HIPER Ring
Advantages
• Predictable recovery times
<200ms / 100 switches
• Simple to implement
• Up to 200 switches in a ring
Disadvantages
• Proprietary
• Only tolerates a single fault
HIPER Ring Q & A (1)
Question:
• Who invented HIPER Ring?
Answer:
• The patent for HIPER Ring technology is jointly owned
by Hirschmann and Siemens
HIPER Ring Q & A (2)
Question:
• When should I recommend HIPER Ring?
Answers:
• When the application requires consistent recovery times
• When the customer wants a clear network topology
• When the geography is not suitable for a meshed structure
• When the customer requires minimized commissioning and maintenance
effort
HIPER Ring Q & A (3)
Question:
• Can I use non-Hirschmann switches in the HIPER Ring?
Answers:
• Yes – if the other switches support HIPER Ring
– Siemens
– OEMs
• No – if the other switches do not support HIPER Ring
– Not possible to reset Learned Address Tables – unpredictable recovery
– Multicast watchdog packets will be broadcast to all end devices
HIPER Ring Q & A (4)
Question:
• Do the settings of the product dip switches affect the
redundancy mechanisms?
Answer:
• Yes:
– Hardware configuration – fixed ports
– Software configuration – free choice of ports
– RSTP enabled / disabled
Check the Installation Guides for product-specific
settings
Fast HIPER Ring
Objective
• Creation of resilient ring structure
Fast HIPER Ring
Advantages
• Predictable recovery times
– 10ms with 10 switches
– 40ms with 100 switches
– 60ms with 200 switches
• Theoretical limit of 20,460 switches in a ring
Disadvantages
• Proprietary
• Only tolerates a single fault
Fast HIPER Ring Q & A (1)
Question:
• Is Fast HIPER Ring supported in all Hirschmann
products?
Answer:
• No. Fast HIPER Ring is supported by the RSR and
MACH1000.
Fast HIPER Ring Q & A (2)
Question:
• Can I upgrade my existing Hirschmann products to
use Fast HIPER Ring?
Answer:
• No. Fast HIPER Ring requires an ASIC* (chip).
* Application-Specific Integrated Circuit
Media Redundancy Protocol (MRP)
Objective
• Creation of a resilient ring structure
HIRSCHMANN HIRSCHMANN HIRSCHMANN
HIRSCHMANN
Media Redundancy Protocol (MRP)
Advantages
• Standardised
– IEC62439
• Predictable recovery times
– Worst case 200ms
• Up to 50 switches in a ring
Disadvantages
• Only tolerates a single fault
MRP Q & A (1)
Question:
• Who invented MRP?
Answers:
• MRP is a standardized version of HIPER Ring, with
some optimizations.
• Any manufacturer can implement MRP in its products
MRP Q & A (2)
Question:
• Are there multiple versions of MRP?
Answers:
• No, but...
• Hirschmann has added some additional features to its
implementation
– Support for ring coupling
– Maximum 200 switches in a ring (standard 50 switches)
– 80ms recovery time (standard 200ms)
• Fast MRP
– Maximum 50 switches
– 30ms/50 switches
– 10ms/15 switches
Start-up MRP
Automatic Ring Configuration & Diagnostic (ARC)
Connect all “out of the box” switches to a ring
Only the “Master” switch needs IP-Parameters
Run automatic mode
Configuration automatically saved in all devices
Multiple Rings
The Sub-Ring concept enables to
easily couple new network segments to
suitable devices in existing redundancy
rings (basis ring). The devices of the
basis ring to which the new Sub-Ring is
being coupled are known as Sub- Ring
Managers (SRM).
Multiple Rings
RM SRM
SRM
Base Ring
SRM
SRM
RM – Redundancy Manager
SRM – Sub-ring Manager
Multiple Rings Q & A (1)
Question:
• Which protocols can be used to create the base ring?
Answer:
• HIPER Ring
• Fast HIPER Ring
• MRP
Question:
• Can all devices be a sub-ring manager?
• Answer:
• Only MACH1000, RSR, MACH4000 and PowerMICE
Multiple Rings Q & A (2)
Question:
• How many subrings can by supported by a SRM?
Answer:
• Four
Question:
• Do all devices in the sub-ring need to support the sub-
ring protocol?
Answer
• No. Only the SRMs need to support the protocol
• The other devices must support MRP
Ring Extension: Ring Coupling
Redundant connection of
Multiple rings
This allows the redundant
coupling of redundant rings
and network segments. Two
rings/network segments are
connected via two separate
paths.
Redundant Ring Coupling
Advantages
• Predictable recovery times
– Average 250ms
• Simple to implement
• Compensates for two faults in a HIPER Ring
Disadvantages
• Proprietary
Redundant Ring Coupling Q & A (1)
Question:
• How does Redundant Ring Coupling compensate for
multiple link failures?
Answer:
• Extended Mode.
HIRSCHMANN HIRSCHMANN
HIRSCHMANN HIRSCHMANN
Redundant Ring Coupling Q & A (2)
Question:
• How can I implement
security functionality
between the two rings?
Answer:
• Use the EAGLE Layer 2
redundancy function.
HIRSCHMANN HIRSCHMANN
HIRSCHMANN HIRSCHMANN
Link Aggregation
Objective
• Create a single high-bandwidth logical link from
multiple lower-bandwidth physical links (trunking)
Link
Aggregation
(LACP)
Link Aggregation
Advantages
• Standardized 802.1ad
• Fast recovery times
Disadvantages
• Short but unpredictable recovery times
Link Aggregation Q & A (1)
Question:
• Can I use Link Aggregation together with HIPER Ring?
Answer:
• Yes, with the PowerMICE and MACH4000
HIRSCHMANN HIRSCHMANN HIRSCHMANN HIRSCHMANN HIRSCHMANN
HIRSCHMANN HIRSCHMANN HIRSCHMANN HIRSCHMANN HIRSCHMANN HIRSCHMANN HIRSCHMANN HIRSCHMANN HIRSCHMANN HIRSCHMANN
IEC 62439
IEC 62439 defines different redundancy protocols:
RSTP, MRP, PRP/HSR, CRP, BRP, DRP
Comparison of Redundancy Protocols
defined in IEC62439
Available
Protocol Standard Typical re-config time Remark since
any topology / mesh,
STP Spanning Tree IEEE802.1 30s 1990
diameter limited
any topology / mesh,
RSTP Rapid Spanning Tree IEEE802.1 2s 2001 / 2004
diameter limited
1 s worst case for 512 any topology / duplicated
CRP Cross Network Protocol IEC 62439-2008
end nodes networks
4,8 ms worst case for Two top switches with star,
BRP Beacon Redundancy Protocol IEC 62439-2008 -
500 end nodes line or ring
100 ms worst case for 50
DRP Distributed Redundancy Protocol Future release of IEC 62439 ring, double ring -
switches
1998 (1) /
MRP Media Redundancy Protocol IEC 62439-2008 200ms 50 switches ring
2007
30ms (50 switches)
Fast MRP Media Redundancy Protocol IEC 62439-2 (2010) ring 2007(2)
10ms (15 switches)
any topology / duplicated
PRP Parallel Redundancy Protocol IEC 62439-2008 0ms 2008
networks
limited to special config
Optimized RSTP Rapid Spanning Tree IEC 62439-1 (2010) 5..20ms per hop 2008
set
HSR High Available Seamless Ring IEC 62439-3 (2010) 0ms ring -
(1) pre-standard Hiper Ring since 1998, MRP since 2007
(2) pre-standard Fast Hiper Ring since 2007
PRP (Parallel Redundancy Protocol)
Zero failover – duplicated networks DAN - Dual attached node
implementing PRP
SAN - Single attached nodes
SAN SAN
SAN SAN
SAN SAN
PRP (Parallel Redundancy Protocol)
Zero failover – duplicated networks DAN - Dual attached node
implementing PRP
SAN - Single attached nodes
DAN DAN
SAN SAN
SAN SAN
SAN SAN
PRP (Parallel Redundancy Protocol)
Zero failover – duplicated networks DAN - Dual attached node
implementing PRP
SAN - Single attached nodes
SAN SAN
SAN SAN
SAN SAN
PRP (Parallel Redundancy Protocol)
Zero failover – duplicated networks DAN - Dual attached node
implementing PRP
SAN - Single attached nodes
SAN SAN
SAN SAN
SAN SAN
Outlook: HSR (High Available Seamless Ring)
Zero failover Redundant Ring
Outlook: HSR (High Available Seamless Ring)
Zero failover Redundant Ring
