Networking the End-to-End Cloud Environment
Emerging cloud offerings from major networking vendors seek to address the complexities of end-to-end cloud data environments.
Everyone wants to be in the cloud these days, but not just any cloud. Enterprises are increasingly looking for a unified, integrated end-to-end cloud that is both flexible enough to handle changing workload demands yet solid enough to provide high reliability and security.
To date, most cloud operations have been patched together from various physical, virtual and software platforms, often from multiple vendors, anchored by leading public service providers like Amazon and Microsoft. In the past few weeks, however, the IT industry has seen a steady stream of integrated hybrid platforms that purport to provide everything from basic provisioning and migration to upper level resource and service orchestration, literally providing a one-stop shop for all things cloud.
Naturally, networking is a major factor in this effort, which is why Cisco launched the Digital Network Architecture (DNA) this week. The DNA portfolio encapsulates everything from local and wide area switching to virtualization and cloud service management, all aimed at building cloud ecosystems that feature broad service and resource automation, in-depth analytics, unified policy orchestration and open, extensible programmability. In short, you get a turnkey global cloud solution capable of supporting the transition to digital, app-driven business models. A key component is the APIC-Enterprise Module (APIC EM) platform, a revised version of the Cisco enterprise controller that provides built-in support for the myriad services and functions needed to support a unified, end-to-end cloud fabric.
Cisco is not the only major IT vendor to draw a bead on the fully integrated cloud. China’s Huawei recently unveiled the Cloud Data Center that combines server, storage and software-defined fabric technologies into a broad ecosystem capable of supporting the entire data stack in the cloud. Huawei has teamed up with key industry players like SAP and Accenture for functions like migration and application development, with broad support for open standards such as OpenStack, Hadoop and Spark, all of which is intended to provide an easy transition for enterprises, mobile operators and telcos to an all-cloud footprint. The system is broad enough, in fact, to provide full hosting of carrier service environments.
Oracle has jumped into the game as well, coming out with an Enhanced Data Rate InfiniBand fabric designed to provide a scalable, non-blocking unified network infrastructure capable of on-demand provisioning and orchestration across the cloud. The platform provides 100 Gbps to enable 4:1 network consolidation as well as integration into legacy data environments, including third-party and customized, home-grown applications. Oracle’s aim is to support the rapid deployment of business-critical apps in local and distributed cloud environments by integrating software-layer functionality with InfiniBand-based fabric processing. The platform will be available by in the second quarter.
Open source champions like Red Hat are gearing up their end-to-end cloud capabilities as well. The company recently joined forces with M2M specialist Eurotech to build an integrated IoT-facing infrastructure that unifies everything from the network edge to back-end cloud services. The system will incorporate Red Hat’s Enterprise Linux and JBoss middleware with the Eurotech Everywhere Software Framework and Everywhere Cloud to put data integration and routing closer to IoT devices in the field. This idea is to provide rapid turnaround of IoT-driven data without having to ship it to centralized facilities for real-time processing and then push it back out again. The solution promises better security, management and application support by embedding key functions on the data, device and application levels.
Most clouds are designed from the ground up to provide operational simplicity, but this often masks the system and infrastructure complexity that accompanies distributed data architectures. By enabling an end-to-end cloud environment, top IT vendors are attempting to simplify the infrastructure side in the hopes that this will lead to even greater operational efficiency.
The chief drawback to these wholly owned or even turnkey vendor solutions is the way they engender over-reliance on a single vendor to support cloud infrastructure. No one can provide all things to all enterprises. So while ease-of-deployment and unified management are important, so too is the ability to incorporate a range of capabilities to suit specific workloads.
This doesn’t necessarily require an open system, but it does call for a platform that can integrate new systems without a lot of fuss.
Arthur Cole covers networking and the data center for IT Business Edge. He has served as editor of numerous publications covering everything from audio/video production and distribution, multimedia and the Internet to video gaming.