Cisco Cisco HyperFlex HX240c M4 Node Leaflet
© 2016 Cisco Systems, Inc. All rights reserved. This document is Cisco Public Information.
Solution Brief
March 2016
Therefore you must be able to optimize the relationship among computing,
networking, and storage resources to support the needs of different applications.
Traditional virtualization clusters completely separate computing and storage
resources, requiring complex SAN technology and costly enterprise storage
systems. Web-scale workloads employ servers with local disk storage using
application software that is infrastructure aware and supports resilience with a fail-
in-place model.
networking, and storage resources to support the needs of different applications.
Traditional virtualization clusters completely separate computing and storage
resources, requiring complex SAN technology and costly enterprise storage
systems. Web-scale workloads employ servers with local disk storage using
application software that is infrastructure aware and supports resilience with a fail-
in-place model.
Existing infrastructure models fail to meet the everyday needs of IT organizations.
The cost and complexity of virtualized environments make them less effective than
they would otherwise be in supporting business applications. The lack of built-in,
application-level resilience of most enterprise applications puts the web-scale
model out of reach.
The cost and complexity of virtualized environments make them less effective than
they would otherwise be in supporting business applications. The lack of built-in,
application-level resilience of most enterprise applications puts the web-scale
model out of reach.
First-Generation Hyperconvergence
Hyperconvergence promised a low-cost, easy way to support a wide range of
applications on a scalable, resilient platform with data distributed across the cluster
servers’ local storage. First-generation hyperconverged products included many
compromises that caused them to fall short of the promise. For example, they have:
•
applications on a scalable, resilient platform with data distributed across the cluster
servers’ local storage. First-generation hyperconverged products included many
compromises that caused them to fall short of the promise. For example, they have:
•
Inefficient scaling: Most products were based on an appliance model that scales
clusters only in fixed ratios of computing and storage resources, not in ratios
tuned to meet the unique needs of applications.
clusters only in fixed ratios of computing and storage resources, not in ratios
tuned to meet the unique needs of applications.
•
Insufficient data optimization: Many products are based on file systems that
weren’t designed to reduce write response times and increase performance of
spinning disks. They typically lack, enterprise-class data services such as data
deduplication and compression, fast, space-efficient clones and snapshots, and
thin provisioning.
weren’t designed to reduce write response times and increase performance of
spinning disks. They typically lack, enterprise-class data services such as data
deduplication and compression, fast, space-efficient clones and snapshots, and
thin provisioning.
•
Narrow workload support: First-generation solutions supported a limited range of
hypervisors, with no plan to address the broader range of application requirements,
hypervisors, with no plan to address the broader range of application requirements,
Highlights
We Define Next-Generation
Hyperconvergence
• We outline goals for the next
generation of hyperconverged
systems so that they overcome the
shortcomings of today’s offerings.
Requirements
• Interoperability
• Hybrid cloud support
• Automated data optimization
• Broad workload support
• Complete infrastructure convergence
• Policy-based security
• Flexible and granular scaling
• Hybrid cloud support
• Automated data optimization
• Broad workload support
• Complete infrastructure convergence
• Policy-based security
• Flexible and granular scaling