Cisco Cisco HyperFlex HX240c M4 Node Leaflet
Criteria for Next-Generation
Hyperconvergence
© 2016 Cisco Systems, Inc. All rights reserved. This document is Cisco Public Information.
Page 2 of 5
such as the needs of containerized
and bare-metal workloads.
and bare-metal workloads.
•
Performance shortcomings: The
network is essential to cluster and
application performance, but it was
left as an unspecified, manually
configured, do-it-yourself project.
network is essential to cluster and
application performance, but it was
left as an unspecified, manually
configured, do-it-yourself project.
•
New management silos: New GUIs
simplified deployment and operation
of cluster nodes, but they added
new tools that didn’t fit in with
existing data center best practices.
These tools lacked capabilities such
as automated server management
and APIs to support programmable
infrastructure and integration
with higher-level tools. Today’s
DevOps environments need these
capabilities.
simplified deployment and operation
of cluster nodes, but they added
new tools that didn’t fit in with
existing data center best practices.
These tools lacked capabilities such
as automated server management
and APIs to support programmable
infrastructure and integration
with higher-level tools. Today’s
DevOps environments need these
capabilities.
•
Security risks: Hyperconverged
environments are dynamic, but they
sometimes sacrifice security to
quickly move virtual machines from
server to server. Network security is
difficult to enforce because virtual
networks are treated differently than
physical ones.
environments are dynamic, but they
sometimes sacrifice security to
quickly move virtual machines from
server to server. Network security is
difficult to enforce because virtual
networks are treated differently than
physical ones.
Defining Next-Generation
Hyperconvergence
A lack of clarity about what IT
organizations really need from
hyperconverged infrastructure has
made hyperconvergence difficult
to define as well. Numerous
companies may call their products
“hyperconverged,” but these
offerings all have different features
and shortcomings that make them
impossible to compare. We propose
a definition of next-generation
hyperconvergence that addresses
these shortcomings (Table 1).
organizations really need from
hyperconverged infrastructure has
made hyperconvergence difficult
to define as well. Numerous
companies may call their products
“hyperconverged,” but these
offerings all have different features
and shortcomings that make them
impossible to compare. We propose
a definition of next-generation
hyperconvergence that addresses
these shortcomings (Table 1).
Characteristic
First Generation
Next Generation
Interoperability
• Creation of new
management islands
• Isolated data not managed
by data center best
practices
• Isolated infrastructure
• No interoperability with
• No interoperability with
other clusters or clouds
• Single point of management
• Consistent policy management
• Consistent policy management
across computing, networking,
and storage resources to reduce
security risk
• Integration with data center best
practices and existing tools
• Integration with hybrid cloud and
support for public cloud storage
• Open API that enables integration
with higher-level tools and
provides programmability
Hybrid cloud
support
• Help in creating private
clouds
• Integration with hybrid cloud
solutions
Data
optimization
• Features, if available, built
as add-ons
• Integrated, always-on enterprise
storage features
• Data lifecycle management
Workload
support
• Virtualized workloads only
• Virtualized workloads with broad
hypervisor support
• Containerized workloads to
support lightweight services
• Bare-metal workloads running
directly on nodes
Infrastructure
convergence
• Software-defined storage
• Software-defined computing with
composable infrastructure
• Software-defined networking
• Software-defined storage
• Software-defined storage
Security
• Virtual networks with limited
visibility and control
• Automated and policy-based
• Isolated application tiers,
• Isolated application tiers,
application instances, and tenants
• Microsegmentation to provide
enhanced security for east-west
traffic within the data center
• Incorporation of physical servers
and virtual machines with
equivalent visibility and control
Scaling
• Rigid, monolithic appliances
• Microscaling of all resources on a
highly granular basis
Table 1. Requirements for Next-Generation Hyperconvergence