DISASTER RECOVERY AND NETWORK REDUNDANCY WHITE PAPER Disasters or accidents would cause great harm on network infrastructure. It is unavoidable and the operation of network would be hampered for a long time. Therefore, the measures on preventing network failure under disasters or accidents are imperative.
Disaster Recovery and Network Redundancy 1. Introduction No matter how well maintained a network infrastructure is, fiber and broadband cabling running through public spaces can, and at any time, be damaged and rendered temporarily inoperable by accidents or disaster. Altai enables businesses to avoid losing real-time access to their remote data centre or Internet outlet in any event by providing a redundant link that will ensure continuity of data transfer to remote disaster recovery site or Internet connectivity. In this paper, the following topics will be covered: What are the common causes of network failure? What causes these failures to occur? Is your network disaster recovery planning good enough to protect damages from any of the above causes? What is the network architecture of disaster recovery solution from Altai? What are the benefits of using wireless solution as network redundancy and what are the advantages in using Altai solution? Can a backup network be used as a production network? What are the differences in network design? 2. Common Causes of Network Failure A typical network structure of an enterprise or campus network is shown in Figure 1. The physical network can be divided into 4 parts and the corresponding possible causes of damages can be grouped into 4 categories which are described below. Figure 1: Typical Network Structure and Causes of Network Failure In-building cable damage in-house cabling ranges from patch cords, horizontal wiring to vertical block wiring. Temporarily damages are usually caused by disconnection of patch cords in various patch panels at floor level
IDF (Intermediate Distribution Frame) or at central MDF (Main Distribution Frame) Permanent damages are usually caused by accidental cut of cables during renovation or in building fire etc. Equipment failure equipment includes router, Ethernet switches or terminating equipment from ISP (Internet Service Provider) such as ADSL modem etc. Failure may be temporarily due to hanging or malfunction of equipment caused by software or configuration issues, or permanently due to equipment failure from a variety of causes such as building fire. Last mile fiber breakage temporarily failures can be caused by faulty terminating equipment such as multiplexer etc, or permanently damages caused by accidental breakage of fiber during civil works, or complete damage of the roads, bridges, trenches, fibers and conduits embedded inside in the event of campus fires, floods or earthquakes. Internet backhaul outage even though all the in-house cabling and equipment is healthy, the network may still go down due to extreme congestion or network outage caused by ISP. This may be due to virus attack which creates tremendous amount of traffic, or due to POP (Point-of-Presence) equipment failure or Internet backbone failure. The backup circuit is usually not large enough to handle sudden surge of traffic routed from damaged circuit. What causes network failure to occur and with what frequency? Here we take a look at a few possible regions and what kind of failure we could expect: Tropical Region Flooding equipment failure, cable damage and Internet backhaul outage due to water damage. Lightning and power outage temporary and permanent equipment failure due to power surge and Internet backhaul outage due to power related issues. Fault Line Region Earthquake in-building cable damage, equipment failure, last mile fiber breakage and Internet backhaul outage due to damage from earthquake. Fire cable damage, equipment failure and last mile fiber breakage caused by fire due to earthquake. Island or Seaside Region Flooding equipment failure, cable damage, and Internet backhaul outage due to water damage. Power outage temporary equipment failure due to power surge and Internet backhaul outage due to power related issues. Typhoon and hurricane strong wind can damage above ground cabling and cause last mile fiber breakage as well as Internet backhaul outage.
Cold Region Snowstorm last mile fiber breakage and Internet backhaul outage due to snowstorm damaging above ground cabling. Power outage temporary equipment failure as a result of power interruption and Internet backhaul outage due to power related issues. 3. Is Your Network Resilience Enough? In order to improve network continuity, IT managers usually think of building a 1+1 redundancy on hardware equipment which may be very costly if the scale of network is large. They first think of equipment redundancy because this is the most obvious item that they can control by themselves. However, the network can still be down due to the other categories of failure as shown in Figure 2 below which they cannot control. Figure 2: Legacy Terrestrial Backup Cannot Fully Guarantee Business Continuity Practically speaking, people seldom provide redundancy on horizontal wiring because it is not worthwhile to double the cost of the whole horizontal wiring portion, when the impact of failure is usually on a single or a few users. In order to lower the risk of last mile breakage and Internet backhaul outage, IT managers of medium size enterprise will connect a second Internet line or leased circuit from another ISP different from the existing ISP, or from the same ISP but from other POP. This arrangement can substantially lower the risk of Internet outage but is not perfect. This is because in most cases different fibers will share the same conduit in the last 500 m or kilometer before entering a building. When accident or disaster occurs both fibers will be broken. On the other hand, to build a complete separate path is prohibitive expensive.
In conclusion, legacy terrestrial network redundancy design cannot fully resolve the service outage problem caused by in-house cable damage and last mile fiber breakage, as may happen during disasters or civil works. 4. Disaster Recovery Network from Altai To provide a disaster recovery backup network that can fully reduce the risks of damages due to in-housing cabling, equipment failure, last mile fiber breakage and Internet backhaul outage, Altai has designed a wireless disaster recovery solution, using its renowned A8n Super WiFi base stations and A2 WiFi Access Points/Bridge as depicted in Figure 3 below. Figure 3: Disaster Recovery Backup Network Architecture
5. Advantages of Altai Backup Network Solution Obviously, the wireless backup solution has many advantages over terrestrial backup solution: Redundancy on access connection while legacy terrestrial backup solution usually does not provide backup to horizontal wiring, Altai solution provides redundancy on access connection not only to building level but down to user level. Wireless is non-terrestrial as Altai backup solution is designed to completely bypass the terrestrial network including both in-house cabling and last mile connection, this means that the Altai backup network is completely insulated from breaks in optical fiber and copper networks which are often caused by civil works and other unexpected activities. Remote sites are less likely to be damaged as both wireless access and backhaul equipment are installed at a distance away from the service building and are usually installed at rooftop in most cases, they are much less likely to be damaged in cases such as building fires, floods, road damage etc. Saves redundancy equipment cost in case of Ethernet switch or router failure, the backup network allows users to bypass the equipment and connect directly and wirelessly to alternative backhaul and therefore a 1+1 redundancy on switching and router equipment is not required. Saves fiber upfront cost laying a secondary fiber for backup in separate trench and routing is prohibitive expensive. The costs are not the glass itself but the costs of laying the fiber especially the cost of trenching or other methods of protecting the strands. The cost of laying a fiber ranges from US$3,000 to US$100,000 whereas the installation cost for a base station can be just a thousand. Rapid deployment wireless backup connections can be done just within hours, whereas laying an optical fiber takes anywhere from 6 weeks to 6 months. Ease of relocation as networking structure is always subject to change, especially for temporary locations or enterprises that relocate frequently, Altai backup solution can take the advantage of wireless and hence relocations are simple and inexpensive. Scalability backup connections can be started from the most critical path and extended to other paths as required at ease. Multiple backups can be upgraded easily.
6. Backup as a Production Network In realty, a backup wireless network can be used as a production network as well, especially for those campuses that do not provide outdoor WiFi service right now. In this case, the coverage design is not limited to minimum service in terms of coverage, throughput and service duration as in case of a pure backup solution; but instead, the coverage can be designed for double or multiple-coverage as shown in Figure 4 below. Figure 4: The Backup Solution Itself Provides Redundancy in Access When access through one base station is failed, the client can re-associate to another base station for continued linkage this means the network can be designed for redundancy in access connection. Further, the wireless backhaul of this backup network can also be designed for redundancy as well as shown in Figure 5 below. When backhaul connection through the default backhaul is failed, the base station can automatically switch to another healthy backhaul, and after the failed backhaul has recovered, the base station can automatically switch back to the default one. The backhaul resilience feature of the base station will continuously monitor the status of the backhaul connection. If anything goes wrong, whether it is from radio failure, wireless backhaul failure, cable damage, LAN equipment failure, last mile cabling damage or Internet outage, the base station will switch to another healthy backhaul. As seen from Figure 5, besides acting as a redundancy network, the backup network can actually be used to off-load normal traffic from the original network. The secondary backhaul can be planned for larger bandwidth which depends on how much traffic you want to direct from the normal traffic.
Figure 5: Multiple Backhaul Structure Can Improve Resilience & Off-load Normal Traffic 7. Further Remarks in Disaster Recovery Planning Satellite or HSPA backhaul in some special areas where Internet backhaul or leased circuit is not easily available, Altai A8n supports backhaul connection to satellite base station or 3G network. Bandwidth control & QoS setting the Altai A8n Super WiFi backup solution provides bandwidth control feature to limit the upload and download throughput per user, base station or SSID such that no single user will consume most of the bandwidth. The A8n also provides QoS setting which allows VoIP application to be set at priority.
8. Ensures Network Connectivity in Disasters With good network redundancy planning, the following summarizes how well Altai solution can provide connectivity in different disaster scenarios: Flooding since Altai equipment is usually placed high above ground, it is not affected by flooding at all. With redundant power backup, it can operate even if an area below it is flooded and short of power. Lightning and power outage with redundant power backup from alternative power source, Altai equipment can keep on working in the event of power outage. All Altai equipment is designed with lightning protection circuitry and is unlikely to be damaged by power related incidents. Earthquake since Altai equipment can be placed outdoor, it is not affected by building collapse or cable breakage. With satellite or alternate backhaul, it can continue to provide connectivity even when all cables have been severed. Fire since Altai equipment can provide redundant coverage, the loss of one or more Altai equipment due to fire can still allow connectivity. With A8n s wide coverage, it is very easy to design and implement redundant WiFi networks that provide backup links. Typhoon and Hurricane Altai equipment is designed to survive strong winds, with very light antenna design and strong mounting solutions; it is unlikely to be damaged by strong winds. With A8n s wide coverage, it is easy to design and implement redundant WiFi networks that provide backup links. Contacts Information: Headquarters: Altai Technologies Limited Unit 209, 2/F, East Wing, Lakeside 2, 10 Science Park West Avenue, Hong Kong Science Park, Shatin, Hong Kong Web: www.altaitechnologies.com Tel: + 852 3758 6000 Fax: + 852 2607 4021 Email: info@altaitechnologies.com Feb 24, 2009 Revised on Aug 16, 2009 Revised on Aug 21, 2013