0% found this document useful (0 votes)
64 views6 pages

Cloud Infrastructure v4

Best Practices for Cloud Infrastructure

Uploaded by

Rodolfo Muñoz
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
64 views6 pages

Cloud Infrastructure v4

Best Practices for Cloud Infrastructure

Uploaded by

Rodolfo Muñoz
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 6

Cloud Infrastructure

Best Practices for Cloud Infrastructure


The data center is changing. As next generation technologies continue to be adopted by organizations looking to modernize or
consolidate their data centers and scale and simplify cloud infrastructure deployments to keep up with the demands of the in-
creasingly digital business. Successful expansion of enterprise infrastructure in smart, safe and modular fashion can be a complex
challenge, while at the same time systems need to allow for future growth and directly contribute to reducing operational and
capital expenses.
Modular, build-as-you-grow infrastructure that provides high availability and high performance in a cloud hosted infrastructure
continues putting pressure on CIOs, IT architects and other key decision makers. The market is flooded with options as hardware
platforms and new solutions proliferate, even as IT budgets flatten or turn downward. Whether you’ve been an enterprise archi-
tect for decades, or your IT department was just asked to implement a cloud hosting system, there’s a morass of technologies to
wade through to find the right components.
Pivot3 pre-configures hyperconverged infrastructure platforms to provide a clean way to bring pre-packaged balanced server
and storage ingredients to cloud deployments that enable quick and flexible modules. These modules provide buy/build as data
center scaling needs expand. Unlike other hyperconverged offerings, Pivot3 preserves the unified storage space and operates as a
true virtual SAN. That means no mismatched storage and compute problems; no high availability trade-offs for capacity efficiency;
and flexible options for capacity using hard disk drives and flash or all flash for high performance.

Architecture Optimized for Performance


Pivot3’s unique architecture is optimized for performance and offers the quality you expect from an appliance-oriented delivery
model. Many of Pivot3’s key differentiators are designed into its operating system, some of which include:
Pre-Validated – Getting the combinations working correctly shouldn’t be an on-site customer experiment. Our engineers have
spent their careers building storage infrastructure. We carefully architect and validate our offerings based on industry-trusted plat-
forms. While standard hardware is used (no custom hardware cards), the combinations still matter and we ensure they perform to
our high standards before the Pivot3 name goes on.
Optimized – Getting the best combination of value comes from our extremely efficient vSTAC platform that enables high avail-
ability and extremely efficient use of storage and compute so you get the best combination of convergence and useable capacity.
After all, if you get hyper-converged and there’s little left for the workload, then that lowers the value. We consume a limited 4
vCPU (virtual CPUs) per server for less than 10% overhead. Our direct IO channel to the storage controller improves throughput
by 30-40% versus alternatives that must traverse the hypervisor; and our storage efficiency reaches up to 94% of useable capacity
while retaining high availability because we do not use an inefficient replication model for data protection.
Built to Scale – A key benefit to using the Pivot3 scale-out option is getting what you need, when you need it and being able to
add to compute and storage as your needs grow. No complex data re-organization is required. Simply add the next node and we
take care of the rest, automatically. You just decide what new use to apply to the extra capacity.
No Hot Spots – Our patented architecture leverages all media on all drives on all nodes. Since the data is written/read across the
array, it’s automatically load balanced. This also means VM activity on any node has the full service IO capacity across the array.
Our direct IO path to the storage controllers, copy-less networking protocol, and full active-active IO path translates into a high
IOPS and low latency system (and without the expense of a Fiber Channel SAN).
Data Protection – Hyperconvergence without data safety is NOT an option. Pivot3’s key design points include protecting users
from themselves, making room for hardware failures, ensuring data and drive integrity, rapid recovery from failures, and pre-emp-

1 of 6
© 2016 Pivot3, Inc. All rights reserved. Pivot3 reserves the right to make changes without further notice to any products herein. The content provided is as is and without express or implied warranties of any kind.
Cloud Infrastructure

tive problem intervention. Our customers depend on five-nines, 24 x 7 uptime on PETABYTES of storage in thousands of deployed
appliances worldwide. Because we always look at the total quality of service coming from the storage system, Pivot3 gives you
these safety capabilities in our scale-out architecture:
• Constant media scanning – On-going surface and performance scanning occurs constantly in the background.
• Drive sparing – We use patented Virtual Global Sparing that gets provisioned by default for a drive’s worth of capacity to
step in should a fault occur, so no downtime or performance degradation.
• Predictable performance– Pre-emptive alerts flag and ‘fail-out’ drives that aren’t working properly and may be affecting
the overall array performance. Our Virtual Sparing steps in and automatically reallocates data to the spare space and alerts
administrators.
• Read/Write Verification – The Pivot3 vSTAC handles all read/write IO and distribution across media in the system. All read/
write verifications occur in a parallel process from each of the system nodes.
• Read Quality Detection – Subtle performance problems arising from a drive getting “sick” may not yet trigger a full failure,
but we constantly evaluate every drive. We also automatically redirect reads to alternative drives to sustain performance
and inform administrators of the drive being in a degraded state. It can help extend the life of a sick drive by lessening the
load to enable a more graceful service event.
• Phone home – Sends support notification alerts to administrators and Pivot3 technicians when there’s a problem needing
attention.
High Availability – We know data loss is NOT an option and while it might sound strange, we design for failure. That means
we don’t consider fault tolerance a tradeoff for performance or safety. The Pivot3 vSTAC platform uniquely supports up to FIVE
simultaneous drive failures or a complete node failure, PLUS two additional drives failing.
One Size Doesn’t Fit All – We offer a wide range of appliance configurations that vary by CPU capacity, RAM, storage size and
form factor. Most of these can be combined to handle differences in workloads that may be heavier on some portions of the
deployment. For example, a server to host the single instances of VMs hosting infrastructure services like Active Directory, SQL,
or VMware vCenter. That means you can economize on other “nodes” that may host different payloads or to function as more
storage-heavy nodes.

Use-Case Driven Options Enable a Wide Range of Workloads

Flexible Form Factors


Graphics Intensive (GPU Enabled)
(12 drives, 12-128TB) • 1U Rack
• 2U Rack
Compute (CPU options)

Processing and Data Intensive


(12 drives, 12-128TB)
• Blade

Mixable in an Array
• Enables a mixture of
Storage Heavy
(12 drives, 12-128TB) compute power
• Storage cohesin still occurs
Light Workloads
(4 drives, 12-128TB)
Flexible Storage Capacity
• Up to 1.5PBs in a System
Storage (Drive Sizes/Types) & RAM • 12 to 128TBs per Node

2 of 6
© 2016 Pivot3, Inc. All rights reserved. Pivot3 reserves the right to make changes without further notice to any products herein. The content provided is as is and without express or implied warranties of any kind.
Cloud Infrastructure

Effortless Management Capabilities Reduce Complexity and Simplify Management


A good hallmark of craftsmanship is getting good precision. At Pivot3, we believe in giving administrative precision with less
chance of “skill” error, so we’ve designed our systems’ administrative tools to take the complexity out of tasks that could lead to
unintended errors. Node additions, drive replacements, LUN allocations, automated data organization/reorganization, and much
more are all intuitively managed via a tool-tip assisted graphical interface.

PIVOT3 INTEGRATION WITH VSPHERE VCENTER WEB CLIENT PLUGIN

Pivot3 vSphere Web Client Plugin dashboard provides a global view for entire domain monitoring, health of all virtual Protection
Groups (vPGs), nodes and volume, as well as total storage by tier, allocated, and Raw.

Out-of-the-Box and Deployment-Ready


At Pivot3, our objective is to get you as close as possible to an out-of-the-box experience. As veterans of this industry, we know what
it costs to test a deployment-ready configuration, so why not leverage our hundreds of hours of validation work? Rather than trying
to get your DIY combination right, we want you to start with deployment instead of architectural planning and experimentation.

Project Acquire Hardware Software Benchmark


Start Appropriate Configuration Install Test Tune and Iterate Deploy Scale Out

Network
Configuration
Pivot3
Starts here

Months of time spent assembling, installing, and configuring

3 of 6
© 2016 Pivot3, Inc. All rights reserved. Pivot3 reserves the right to make changes without further notice to any products herein. The content provided is as is and without express or implied warranties of any kind.
Cloud Infrastructure

A Storage Efficient and Cost-Effective


Hyperconverged Methodology
Pivot3 offers a better way to do hyperconverged that is more cost
effective than your Fiber Channel SAN. We don’t believe in trading off
ease of use for high-end performance. The P platform doesn’t require
that you trade off data safety for cost-effectiveness. When you need Flash Tier
24 x 7 operations that won’t sacrifice your service level obligations,
then come to a company that knows how to keep your operational
RAM Tier
efficiency and reliability at world-class levels.
At Pivot3, the solution is simple. We bring all media installed on all
nodes together as a unified storage space. As you add nodes, you truly
add more storage. We don’t have islands of media that strand unused
space and our useable capacity protecting the entire array includes
multiple drive HA protection and a patented virtual global spare
that is shared across the array. The Pivot3 operating system actually
increases useable capacity as you add nodes (up to 94% useable Pivot3 implements patented
capacity at 16 nodes). Scalar Erasure Coding to:
• Unify hardware resources: RAM cache,
How All Storage Capacity Becomes Unified in SSD and HDD across nodes
the Pivot3 Platform • Wide Stripes data across all unified
Patented technology turns all drives into a single iSCSI SAN shared hardware
by all nodes. Any VM on any node can access the full pool and the • Data written once and only moved when
aggregated IOPS of all drives. new nodes are added
• Aggregate IOPs and Bandwidth as nodes
are added

Pivot3 HyperSAN
• Unified pool of SSD and HDD resources
• Data is wide-striped in 1MB segments evenly
across the entire virtual
Performance Group
• Pivot 3 storage Capacity presents as an iSCSI
target
• Volumes of any size and quantity can be
provisioned
• Select different Data Protection levels for
any volume

4 of 6
© 2016 Pivot3, Inc. All rights reserved. Pivot3 reserves the right to make changes without further notice to any products herein. The content provided is as is and without express or implied warranties of any kind.
Cloud Infrastructure

Seamless Scalability
As cloud deployments scale to petabytes, Pivot3 enables massive capacity with flexible options through scale-up. All hyper-con-
verged nodes are centrally managed through the vSphere web client plugin for a single operational control panel. The vSphere
web client plugin automates the addition of nodes, LUN volumes, high availability options, and any service events like software
upgrades or drive replacement. All of these events come to you from Pivot3 with “low touch” requirements because we automate
the data mapping process for scenarios like adding another node. Terabytes of expansion occur seamlessly and without admin-
istrative specialization.

vSTAC Domains and Virtual


Performance Groups (vPG)
• Multiple vPG’s achieve
the best combination of
reliability, performance, and
capacity
• Multi-petabyte Scale in a
vSTAC Domain!
• All unified through the Pivot3
vSTAC Management Console

Unlike other software-only options, we prefer to pre-engineer our solutions to avoid critical errors – including unplanned outages.
Consider the important statistic below from 451 Research. They asked IT professionals if they’ve ever had an unplanned storage
outage. Two out of five respondents said they had. Nearly forty-percent of those unplanned outages were due to human error.
Cloud scale deployments require dependability and protection from all avoidable problems.

5 of 6
© 2016 Pivot3, Inc. All rights reserved. Pivot3 reserves the right to make changes without further notice to any products herein. The content provided is as is and without express or implied warranties of any kind.
Cloud Infrastructure

UNPLANNED STORAGE OUTAGES

Unplanned Storage Outage Reasons

Vendor Hardware/Software
72%
Failure

Yes
40%
Human Error 39%
Yes

No Natural/Unnatural Disaster 11%


60%

Vendor Support Patch Failure 8%

Source: 451 Research – Storage Executive Summit

Pivot3: Future-Proofed for the Next Generation


Next generation technologies continue to be adopted as organizations look to modernize their data centers and scale and simplify
cloud infrastructure deployments to keep up with the demands of the increasingly digital business. Hyperconvergence is one of
the fastest-growing methods for deploying IT in the data center as IT departments look for ways to adjust to their new role as
business-enabler and respond to the new demands that are placed on them. From datacenter design to strategies that align
IT with business units, organizations are forced to rethink how they operate, how they manage resources, and what it takes to
increase time-to-value on IT initiatives.
Pivot3 simplifies IT complexity with hyperconverged solutions that enable businesses to dynamically assemble the right IT
environment to efficiently and reliably support digital business. Pivot3 addresses the pressure on IT departments - to instantly
provision resources, rapidly create and deploy software and services to grow revenue - with the agility to get applications up and
running without interruption. Thousands of organizations rely on Pivot3 for mission-critical applications, server virtualization, high
performance storage, private and hybrid cloud environments, as well as general purpose workloads like virtual desktop infrastruc-
ture and remote and branch office, and more.

6 of 6
© 2016 Pivot3, Inc. All rights reserved. Pivot3 reserves the right to make changes without further notice to any products herein. The content provided is as is and without express or implied warranties of any kind.

You might also like