Virtualization and Cloud executives share their predictions for 2017. Read them in this 9th annual VMblog.com series exclusive.
Contributed by Jeff Roper, CTO, Entuity
SDN, SD-WAN and Right Sizing In The Cloud
Despite predictions over the past four years or so of a vast roll out of SDN across enterprise networks, SDN uptake continues to be cautious at best with a refocusing on more pragmatic, less risky endeavours. Given that the drivers behind the deployment of SDN have not changed and neither have the risks and obstacles, it is likely that SDN will continue to be more widely deployed throughout 2017 but at a similar cautious pace. The lack of credible network management products which are capable of effectively monitoring and troubleshooting SDN deployments and also the non-SDN to SDN boundaries continues to impede uptake.
One exception to the slow roll out of software defined networking is the rapid and accelerating deployment of SD-WAN. Whilst SD-WAN could be viewed as rebadged, smart load-balancing or policy based routing, the ease of deployment (as a single virtual or physical appliance), low complexity (compared with full SDN), low entry cost, acceptable risk and obvious business benefits are resulting in a significant uptake in SD-WAN appliances. Given the relentless migration of on-premises applications and storage to the cloud and consequently a greater reliance on internet connectivity from within the enterprise, many enterprises are turning to SD-WAN (deployed in conjunction with lower cost broadband connections) to satisfy their growing internet connectivity needs to avoid increasing the bandwidth of their costly dedicated WAN links. In 2017, we can expect to see the SD-WAN deployment trend continue and, with the SD-WAN market settling and maturing, we can expect to see solutions mature and begin to offer integration points with other parts of the IT infrastructure - most notably network management systems and security solutions.
With increasingly business-critical applications being deployed to the cloud, identification of the probable causes of faults becomes increasingly complex. Traditionally faults fell into three main areas - application, server (physical) or enterprise network. Application management has changed the least, server management has acquired the complexities of virtual resource limitations but it is with the network that fault identification has become much harder - both to diagnose and to resolve quickly. Problem resolution is hampered by difficulties in identifying whether the fault is in the enterprise's own network, with one of their ISPs, elsewhere in the Internet or within the Cloud Service Provider's network (either internally or virtual networking controlled by the enterprise). In 2017 it is predicted that tools will begin to embrace these difficulties helping achieve more rapid problem identification and resolution - especially of soft faults (such as performance degradation) as well as outages.
Given that one of the main drivers, if not the main driver, for migrating applications to the cloud is for cost reduction, it is interesting that research continues to show how many virtual servers in the cloud are over-provisioned. Some over-provisioning is justifiable during the initial migration phase to ensure that applications continue to receive the resources they require, but in 2017 when migrated applications are proven to be performing satisfactorily there will be greater pressure to "right size" virtual machines in order to maximise savings. This requires careful analysis of applications' resource requirements (under both routine and peak loadings) and a greater use of elastic provisioning.
About the Author
Jeff Roper, PhD - Chief Technology Officer
For over 10 years, Dr. Roper's contributions to the overall success of the Entuity Network Management solution have been as far reaching as they are deep. They include projects that both increase the overall understanding of how networks operate while also being translated into solid product features that help ease the overall burden of managing several hundred to thousands of network devices scattered around the globe. These include significant advances and enhancements in the areas of QoS, MPLS, IPSLA, service monitoring, configuration management, flow analysis, topology discovery, AA-NPM, virtualization support, routing protocols, XML API support, device support, device modeling, VoIP, and IPv6.
Prior to joining Entuity in 2002, Dr. Roper held several position of increasing responsibility at ViewGate Networks, an innovator in providing real-time, end-to-end service management, including Senior Design Engineer, Team Leader and Software Architect. Before moving to ViewGate, Dr. Roper held a variety of positions with British defense agencies culminating in the position of Senior Scientific Officer/Senior Scientist with the Defence Evaluation and Research Agency. Dr. Roper received his BSc with Honours in Computer Science from Nottingham University before going on to receive both a Masters (1995) and PhD (1999) from Cranfield University. Dr. Roper is also a Chartered Engineer, a Chartered Information Technology Professional, and member of the British Computer Society.