SEMUA MENGARAH KE HYPERCONVERGENCE

fankychristian - August 01, 2016

Prepare for the Next Phase of Hyperconvergence

22 January 2016 | ID:G00298562
Analyst(s):
 

Summary

HCIS is not a destination, but an evolutionary journey — and we are at the cusp of a new phase. I&O leaders must widen their HCIS lens to encompass continuous application delivery, which will propel HCIS toward mainstream use in the next five years.

Overview

Key Findings

  • We are on the cusp of a third phase of integrated systems: Phase 1 is the peak period of blade systems (2005 to 2015); Phase 2 marked the arrival of converged infrastructures and the advent of hyperconverged integrated system (HCIS) for specific use cases (2010 to 2020); Phase 3 represents continuous application and microservices delivery on HCIS platforms (2016 to 2025).
  • The current second phase (through 2020) is primarily marked by specific, stateful hardware and hypervisor-based appliances, with the advent of decoupled hardware and software stacks for infrastructure as a service (IaaS), and the HCIS market will grow at a 46% compound annual growth rate (CAGR) between 2015 and 2019.
  • The third phase will deliver dynamic, composable, fabric-based infrastructures by also offering modular and disaggregated hardware building blocks, driving continuous application delivery and continuous economic optimization.
  • This evolution of HCIS presents I&O leaders with a framework to evolve implementations and architectures to the upper right of the integrated systems landscape.

Recommendations

  • I&O leaders should add two new search and evaluation criteria to HCIS considerations: continuous application delivery and continuous economic optimization.
  • I&O leaders should map enterprise long-term business objectives to continuous application delivery and continuous economic optimization progression.
  • IT planners and enterprise architects should assess the fluidity with which resources can be aggregated and disaggregated to deliver, add, modify and repurpose applications and microservices.
  • IT planners and enterprise architects should design proofs of concept and validate vendor Phase 3 roadmaps for: continuous application delivery; continuous economic optimization; stateless and stateful applications (for example, hypervisors and containers); composable, Mode 2 DevOps; fabric resource and network management; disaggregation; and continuous service delivery.

Analysis

The current hyperconverged integrated systems market will grow from $371.5 million in 2014, to nearly $5 billion by 2019 (see "Forecast Analysis: Integrated Systems Worldwide 1Q15 Update" ), or a 68% CAGR (see Figure 1), according to Gartner's Forecast Analysis. Despite the high growth rates, HCIS use cases are still limited. While we fully expect the use cases to embrace mission-critical applications in the future, current implementations could still pose constraints on rapid growth toward the end of the decade without further developments as described below.
Figure 1. HCIS Forecast Assumptions
Research image courtesy of Gartner, Inc.
Source: Gartner (January 2016)
Among the further developments in and current constraints to HCIS growth are:
  1. Use-case breadth and coverage for large enterprises have been limited, causing silos with existing infrastructure.
  2. Virtualization is primarily hypervisor-based and stateful, requiring setup and configuration skills.
  3. Most HCIS applications are still regular enterprise Mode 1 applications, which also run on existing infrastructures causing a clash of implementations.
  4. Most solutions have targeted simplicity and ease of setup and deployment, but with predictable workloads and less complex applications.
  5. Fail-fast Mode 2 applications and services (see "Predicts 2015: Bimodal IT Is a Critical Capability for CIOs" ) are still maturing and have not been widely integrated as part of the software stack solutions on HCIS.
  6. On-premises or hybrid cloud solutions are in early phases (for example, OpenStack), while HCIS suppliers prefer to drive the mature environments for quicker sales.
  7. Resource pools are often fixed and rigidly specified, rather than being granular, elastic and composable from disaggregated digital building blocks.
  8. True software-defined networking (SDN) intelligence for complex scale-out and scale-up needs is still largely lacking or in an early phase.
  9. Architectures are primarily biased on scale-out clustered implementations, and remote office/branch office (ROBO) and distributed sites, but performance at large scale is mostly unproven.
  10. Pay per usage and licenses proportioned to resource utilization are currently largely experimental.
However, these limitations can and will likely be overcome in forthcoming generations. To show continued HCIS potential, we have extrapolated and developed an integrated system landscape showing the evolution of HCIS in the broader integrated system context, as graphically depicted in Figure 2. The basis and foundation of the model is derived from a number of sources, listed in the Evidence section.
Figure 2. The Integrated Systems Landscape: Evolution and Expansion
Research image courtesy of Gartner, Inc.
Source: Gartner (January 2016)
The degree of evolution is constructed across two axes. The vertical axis indicates the degree of maturity in attaining continuous application delivery. The fundamental premise of HCIS has been five attributes (see "Five Keys to Creating an Effective Hyperconvergence Strategy" ). From inquiries and focus groups, we have learned that in addition to total cost of ownership (TCO), two of the most important attributes were speed and flexibility. "Continuous" implies both attributes. The faster the time to set up, provision, deploy and enhance configurations and services with quality of service, the quicker the response to market needs and the greater the revenue and profit opportunity (that is, continuous economic optimization — the horizontal axis). While speeding deployment and managing change expeditiously, the entire system must continually adapt to changing resource demands. Continuous economic optimization implies efficient resource utilization in near real time, granular and elastic scaling, disaggregated and inexpensive modules, software-defined implementations, paralleling infrastructure as a service.
The bubbles in Figure 1 represent the degree to which integrated systems (in different phases) have met the optimization requirements of the two axes. Starting with blade solutions in the 2005 to 2015 period (periods are approximate), we derived comparative bubble positions for converged infrastructures in the earlier part of the 2008 to approximate current 2015 to 2016 period. We then added integrated system developments, known as HCIS (and primarily hardware and appliance driven), in the 2013 to current 2016 period. We continued to illustrate the evolution of HCIS from the current period toward the 2020 time frame, as software-based solutions become a driving force, and we project further developments of HCIS (which may assume new nomenclatures, such as microconvergence, composable and fabric-based HCIS microservices platforms).
Whatever names these more advanced versions of HCIS are labeled, it is clear from bubble positioning that they will advance both up and right along the two optimization axes as we move toward 2025. Thus, we're suggesting HCIS development and growth is an evolutionary process — not revolutionary — and its progression will be dependent on multiple hardware and software advances (for example, networking and software-defined enterprises). Ultimately, the underlying infrastructure disappears to become a malleable utility under the control of software intelligence (that is, infrastructure as code), automated to enable IT as a service (ITaaS) to business, consumer, developer and enterprise operations.
Note the terminology contrasts past and current versus future:
  • Past and current: Consolidation, virtualization, reduction, simplification, Mode 1, stateful
  • Future: Hybrid cloud, software-based, continuous, fabric-based, microservices, stateless, Mode 2 agility, composable, power and space saving, added operating expenditure (opex) savings
In future research, we will categorize and highlight vendor solutions using the continuous application delivery/continuous economic optimization model.

Conclusions and Recommendations

  • Use the landscape model to make strategic vendor choices — using tactical solutions as trials.
  • Make vendor choices on the basis of their maturity in driving a continuous application delivery/continuous economic optimization deliverable model.
  • Seek solutions that present resource pools as modular, disaggregated, and composable by intelligent and automated software management, tools and APIs.
  • Architect infrastructure to be a service-driven platform (for example, cloud, templates, recipes, pay as you grow).
  • Prioritize vendors with intent to expand bimodal delivery as part of the HCIS platform.
  • Apply five key determinants (simple, flexible, selective, prescriptive, economic) as part of the evaluation process.

Evidence

The model of integrated system evolution is based on client inquiries, vendor briefings, user roundtables, focus group live chats and Gartner research on HCIS.