In our digitally-driven society, outages of high-profile online services can quickly become breaking news headlines. And if you ever wondered how those outages could possibly happen, check out these insights from the Uptime Institute. Between January 2016 and June 2018, researchers collected data from over 100 major service outages and between February and May 2018 conducted a survey of 1,104 data center operators and IT practitioners globally. Here is what they found: 83% of these outages were caused by failures in power sources (36%), networks (25%) and IT systems (22%).[i]
Needless to say, network redundancy is mission critical when it comes to delivering business application performance and continuity. Software-defined interconnection makes designing and deploying redundant networks easier and faster to achieve.
What is network redundancy and why is it so critical?
You know the old saying, “Two is better than one?” It’s a good analogy for network redundancy. If one network device or link fails, then there is a backup to help prevent service outages. And if you have two networks running in parallel, they can share the load (load balancing), which is also an important consideration for business application performance and continuity. Overloaded networks can increase latency (delays), packet loss and jitter in voice and video applications, and can be the root cause of network failures.
So why don’t more IT organizations deploy network redundancy? Cost is probably the leading deterrent. Duplicating network devices and links can be expensive, but nowhere near what it costs for just a minute of downtime. According to the Uptime survey, nearly a third of outages reported by respondents cost them over $250,000 per incident.
Network redundancy made simple
Another reason why IT administrators may avoid deploying network redundancy is that it introduces complexity into their infrastructures. But that doesn’t have to be the case. With software-defined networking technologies, you can design and deploy redundant networks in minutes through API-driven tools and pre-configured user-friendly portals.
You can also leverage software-defined interconnection to privately connect primary and secondary data center IT infrastructures for fast, secure failover and application recovery. A software-defined interconnection fabric, such as the Equinix Cloud Exchange Fabric™ (ECX Fabric™), provides access to multiple cloud services (either from the same or different cloud providers) across multiple metros, which you can then connect to using your network service provider (NSP) of choice. The diagram below illustrates how you can deploy a hybrid IT infrastructure that provides multiple levels of redundancy at the data center, network service provider (NSP), and cloud service provider (CSP) levels:
- Data center redundancy: Multi-metro interconnection across data centers can be achieved via ECX Fabric. Direct, private interconnection is more reliable and secure than the public internet, which has a high risk of network congestion and failures beyond your control. ECX Fabric acts as a network aggregation and data exchange point for multiple ECX Fabric NSP partners and their customers. ECX Fabric can be ordered and configured with dual ports (dual-homed) for instant, redundant connectivity to two separate NSP links and two ECX Fabric connections. In addition, point-to-point SD-WAN connections can co-exist with ECX Fabric if you wish to segment your data traffic over two different networks for any reason. This flexible design also enables ECX Fabric and SD-WAN infrastructures to run in tandem as backup connection services for each other.
- Network redundancy: Many NSPs around the world leverage IP peering at Equinix, via the Internet Exchange, to scale network bandwidth and deliver the best possible quality of service (QoS) to their customers via low latency connections. In redundant network configurations, customers can select the “best paths” from their NSPs with the lowest latency and highest throughput to balance the data traffic load between source and destination points of presence (PoPs). ECX Fabric enables one-to-many interconnection to NSPs, cloud, SaaS and IT service providers, enterprises and businesses’ infrastructures over high-speed, low-latency virtual connections through its redundant physical ports. This network redundancy enables both load-balancing performance and business continuity in the event of a network failure.
- Cloud redundancy: A completely redundant ECX Fabric network configuration, as shown in the diagram below in SV and DC Equinix Interconnection Business Exchange™ (IBX®) data centers, provides two on-ramps to densely populated cloud ecosystems. Through ECX Fabric, any customer can gain access to these cloud ecosystems and leverage them as primary and secondary cloud environments for greater business continuity.
Again, the key is to build in as much redundancy as possible between data centers, networks and clouds.
Fully Redundant Multi-Metro Architecture Using ECX Fabric
How to quickly and easily set up redundant networks
Here are some tools that will help you design, configure and deploy redundant networks using standard, software-defined APIs and the customer portal for ECX Fabric:
- ECX Fabric Network Link Map: The ECX Fabric network link map is one of the tools we use to help our customers plan their ECX Fabric deployments in metros around the world. It provides real-time statistics on NSP link availability, round trip time, packet loss, network utilization and jitter between ECX Fabric PoPs in metros around the world. This will help you plan out redundant network routes that deliver the best QoS, as well as create primary and secondary network paths for disaster recovery. One important thing to remember when designing your redundant network for either performance or business continuity, is to configure both networks to be able to handle 100% of the load, if needed. That way if there is a spike in performance or a complete failure of one of the networks, you’re covered. And since you can configure these software-defined connections to be billed in daily increments, you don’t have to worry about paying for unused capacity. In addition, you can develop automated scripts so that the secondary network spins up or down as needed.
- Customer Portal: The ECX Fabric Customer Portal helps you configure network and cloud connections to service providers, another Equinix customer or your own infrastructures on the ECX Fabric within Equinix IBX data centers.
In the ECX Customer Portal, you can also provision primary and secondary cloud connectivity over redundant networks in just minutes to achieve automatic failover or load sharing. Redundancy is built in to each ECX Fabric hub and, in some CSP configurations, such as Microsoft Azure, duplicate cloud connections are required.
With software-defined networks and interconnection APIs and tools, network redundancy is well within reach, and well worth the upfront time and investment you put into ensuring that your digital business is running smoothly and continuously.
Start by checking out the Cloud Readiness and Cloud Optimized WAN workshops, delivered by Equinix Professional Services’ network and cloud computing experts. They will help you better assess your current IT infrastructure and develop a winning strategy for designing and deploying redundant networks and hybrid multicloud configurations for greater performance and business continuity.
[i] Uptime Institute, “Uptime Institute data shows outages are common, costly, and preventable,” June 2018.