Not too long ago the industry was abuzz about this new thing called cloud. While the chatter remains high, a lot of us have gotten down to the knitting so to speak.
What differs is that some of us jumped on the innovation bandwagon faster than others. While some vendors repackaged existing technologies for the cloud, others got on to addressing the real crux of the change brought about by cloud computing: how we deliver IT services.
As discussed in previous posts about delivering IT as a service and a new way of thinking about a single pane of glass, what matters most in today’s world is agility and innovation for competitive advantage. If you’re an IT industry vendor, this means you need to provide the ability to quickly and easily deliver new services—not just end-user services, but infrastructure services as well—with an increasingly robust tool set to address a variety of technical and business needs.
The main drivers for cloud computing are agility, cost, and efficiency. In fact, three-quarters (75%) of IT executives cite business agility as the main driver according to CIO magazine’s 2011 “Global Cloud Computing Adoption“ survey.
More than half of the executives in this same survey also cite reducing IT infrastructure investment (56%) and IT management and maintenance (53%) as top drivers too.
As the IT industry has settled into deploying technologies for cloud, three (3) options have emerged for deploying services:
1. Legacy bespoke solutions: ala carte ensembles mostly for specific applications requiring certain technologies or for those customers with the means to afford it.
2. Reference architectures: tested, documented, quantified, and qualified solutions for a variety of deployment needs.
3. Single SKUs: pre-configured hardware and software combinations consumed in blocks of capacity that can be combined to scale out a data center.
Converged infrastructure dominate these deployment options, whether deployed with reference architectures or single SKUs. These preferences make sense if you consider that tried and true approaches like those embodied in reference architectures and single-SKU solutions can be deployed more rapidly, at less expense, and with varying degrees of integration.
What are the similarities?
Most of the major suppliers (e.g. EMC, HP, IBM, and Oracle) offer some flavor of these three options and may even look alike at first glance given the shear breadth and depth of their product and service portfolios. EMC, for example, has standalone solutions, reference architectures (including the new VSPEX Proven Infrastructure), and building blocks (Vblock platforms via the VCE – VMware, Cisco, and EMC coalition).
What are the differences?
Where it differs is in the software layer. Hewlett-Packard has HP Converged Cloud (a combination of HP technologies and the free open source OpenStack) and IBM has SmartCloud (technologies for delivering infrastructure, software, and platform as a service). EMC, on the other hand, has forgone slapping a new label on the same old, same old in favor of some new innovations.
To date, only EMC has a management offering that pulls the pieces together and presents the compute, network, and storage resources as a single, converged infrastructure management service.
When introduced more than two years ago, EMC Unified Infrastructure Manager (UIM) was the first of its kind to harness the power of a single-SKU Vblock and remains a unique piece of technology in how it simplifies the creation and delivery of cloud-based infrastructure services through policy-based automation.
Score: +1 EMC for agility.
EMC UIM transforms Vblock platforms into single, virtual environments. For the uninitiated, this EMC infrastructure management offering comes in two (2) versions: 1. Unified Infrastructure Manager/Provisioning (UIM/P), and 2. Unified Infrastructure Manager/Operations (UIM/O).
Focusing on the provisioning aspect (operations will be covered in a future post), UIM/P is purpose-built for Vblock platforms. Use it to interact with compute, network, and storage element managers to define available resource pools and view capacity through a single dashboard.
Create policy-based, automated converged infrastructure provisioning templates comprised of compute, network, storage and operating systems, with grade levels (think gold, silver, bronze) matched to business application requirements. These templates speed service delivery on any Vblock platform, while reducing the risk of human error. Service offerings are then placed into a catalog, ready to be provisioned with just a few clicks.
Score: +1 EMC for reducing IT management.
UIM/P also offers service lifecycle management with elastic provisioning. In its initial release, UIM/P made it possible to standup new infrastructure services five times faster than without it. Add or remove component resources as needed to support business demand fluctuations; allocate the right resources as needed without incurring additional expense of wasted capacity. Additionally, Vblock platform configuration compliance analysis and tracking ensures error-free provisioning and full compliance with Vblock platform standards.
Score: +1 EMC for reducing infrastructure costs.
While vendors like HP and IBM have yet to match the fundamental capabilities of the first generation EMC release, UIM/P (current version 3.1) has gone on to next-generation capabilities such as a provisioning time 30% faster than the previous release and added provisioning and elasticity support for NFS datastores, a must-have when integrating legacy data storage systems.
Additionally, UIM has been architected with REST APIs built to the EMC Data Access API (EDAA) specification. This approach is in keeping with the cloud model concepts of agility and self-service. REST APIs open up this infrastructure management software to future integrations and extensibility via anticipated mashup technologies for quickly creating custom dashboards.
UIM REST APIs also compliment VMware vCloud Director and vCenter APIs to form the foundation for a cloud service delivery platform that provides seamless end-user service management. These APIs enable workflows and processes that can be executed automatically based on environmental thresholds or at the discretion of an administrator to speed service delivery, opening up many possibilities for IT (or compute) as a service.
Score: bonus points EMC for continued innovation.
There is a wealth of possibilities out there for cloud deployment options including single-SKU. But beware of the simple bundle of silos; those technologies without the benefit of innovation to manage and deliver IT infrastructure services across domains—for agility, lower costs, and more efficient management.