Atchison Frazer – KEMP Technologies CMO
What was striking about Day One of the Gartner Data Center Conference here in Las Vegas is that for what is essentially a networking and infrastructure focused event, with an audience drawn from the large enterprise end-users who build and manage networks, the over-riding theme is that their world is now upside down in that the applications are what drive data center architecture and investment decisions.
According to Gartner VP David Cappucio, in his keynote opening address, what business users and line of business managers really care about are real-time operations (RTO) built upon an Intelligent Data Center whose design is largely dictated by priority business applications, many of which are more easily raised, updated and taken down in a global topology dependent upon more effective integrated orchestration and dynamic workload allocation at the application layer.
Cappucio said that this trend will place more emphasis on the cloud or virtualization infrastructure, and tangentially, the cloud services brokers or SI middlemen who can more efficiently aggregate and allocate cloud resources for those enterprise app workloads that make sense to run in the cloud temporarily or for multi-year periods. And thus, the professional role of “DevOps” becomes much more critical in terms of speeding up the growth of applications and making quick judgment calls relative to what apps need to be maintained for continuity purposes and which do not (upwards of 25% of legacy applications, typically way overprovisioned, are still running in conventional datacenters but serving no real business purpose).
In another keynote interview with Don Duet, VP-Technology at Goldman Sachs, it was evident that for a large financial services institution, the ability to move and change applications in software running on top of a “commodity curve” physical infrastructure is now a business critical priority. Duet said that Goldman is placing four “Big Bets” on the future of its business, which is now effectively an IT business:
- Commodity hardware systems designed and built by Goldman Sachs, with 6,000 software developers worldwide who can move data and applications to respond to rapidly shifting market conditions (read: RTO).
- Software-defined everything; the applications influence the scale-out infrastructure, and business intelligence software determines priorities.
- Drive infrastructure change from an application-centric focus.
- Open architecture to drive IT agility (open standards and open sourced code).
What the IT industry needs is a Moore’s Law for driving down the cost of complexity while driving up strategic value and risk factors to the application layer,” said Duet.
Finally, Gartner analysts Andrew Lerner and Joe Skorupa outlined their vision of the industry’s latest buzzword, SDN or software-defined networking, but the taxonomy belies the real intent of SDN, which is to empower the application and the “DevOps” professional to have the flexibility and operational agility to bridge the conventional gap between applications and the network. Lerner and Skorupa went even further, describing the real opportunity inherent in SDN in terms of an innovation platform – net new applications yet to be conceived and designed.
This is a hugely important paradigm shift when you take into account that there are now over 10,000 APIs published in the Programmable Web directory, spawning over 7,000 new business applications (mashups). Those APIs almost exclusively run in a virtual machine container and thus must be optimized to ensure the highest levels of availability, performance, security and entitlement for reuse in other mashup applications, and cross-domain in terms of technological entities (public-private cloud).
The real value in SDN, in the analysts view, is an intelligent chain of services, including the upper layer of network function virtualization (NFV) services like security and application delivery controllers (ADCs) that are fully integrated with an intelligent, open interface and common communications protocols that for the first time will understand and manage the topology of the network end-to-end.