Oct 2, 2017
The rise of cloud applications has been well documented on this site and others. The cloud era kicked off with a handful of SaaS applications, such as ERP, CRM and HR systems. Today, businesses are buying almost everything from the cloud from compute services, contact center software, unified communications to anything else you can think of. These apps and services may look somewhat unrelated, but they all have one thing in common, they are highly dependent on the network to perform properly.
Consider a consumer example. We pay Netflix a certain amount of money to watch a movie, but we are entirely dependent on our network connections to enjoy the experience. It doesn’t matter how much more money we pay Netflix – if your network connection performs poorly or is unavailable, the viewing experience suffers. Similarly, in the business world, if the network fails, so too will the application. In some cases, it may not matter, but if it’s something like a cloud contact center service, an outage could have an immediate and direct impact on customer service.
Businesses are non-stop organizations that run 24 x 7 x 365 and any application outage is unacceptable. This requires a network that has an availability that is truly “five-nines” across both the LAN and the WAN. In the past, I have found the term five-nines to be more of a marketing term and when I press the vendor using it on whether their solutions truly offer that level of availability, I get a bunch of caveats. Today, there can be no caveats or compromises. Five-nines must mean 99.999% available and nothing less.
An SD-WAN that utilizes dynamic path control ensures that the WAN services and underlying transport is highly available. Instead of having to operate in an ‘active-passive’ mode where failing over could take several minutes, an SD-WAN can leverage multiple active connections and offer sub-second failover.
However, there’s more to a network than just connectivity. Another requirement would be that the hardware, particularly in the branch offices be configured for high availability. The router vendors have had this capability for decades with protocols such as VRRP (virtual router redundancy protocol) or HSRP (hot standby router protocol), but this requires buying two identical routers with one acting as a standby for the other. This approach also requires that each set of routers be manually configured and then the configurations kept in lock step. It also requires redundancy in WAN services, requiring multiple IP addresses from each provider for both the active and standby routers. The complexity and cost has held back most organizations from doing this. A better approach would be to apply lessons learned from the computer industry where clusters can be put in high availability mode, which provides faster scale, higher uptime and simplified management. If networking transforms to software running on off-the-shelf compute platforms, then the industry can apply the same high availability options the server industry has come to rely upon. An advanced SD-WAN solution should offer the ability to configure active/active redundant clusters without requiring redundancy of WAN services and IP addresses. A highly available SD-WAN can provide 99,999% availability of branch infrastructure and services.
In parallel, many organizations are trying to limit the amount of on-premise infrastructure, particularly in space-constrained branch offices, so building a compute cluster may not be ideal. For those businesses, a new option is emerging that leverages the power of the cloud. There aren’t many available, but one interesting solution is the recently announced Unity Orchestrator-as-a-service from Silver Peak. The product gives customers the high availability they require with a zero-capex model by leveraging container instances that are managed via the cloud. This eliminates the need for on-premises compute and storage infrastructure to manage an SD-WAN.
More and more organizations are adopting a cloud-first model, raising the value of the network from being a tactical resource that most executives don’t give a second thought to a strategic asset that embraces the cloud with the right architecture to move their businesses forward. High availability is no longer an option, it’s something organizations of all sizes must consider as part of their digital transformation strategy.
Zeus Kerravala is the founder and principal analyst with ZK Research. He provides a mix of tactical advice to help his clients in the current business climate and with long term strategic advice. Kerravala provides research and advice to the following constituents: End user IT and network managers, vendors of IT hardware, software and services and the financial community looking to invest in the companies that he covers.