Software-Defined Servers explained.

How they work. Why they’re essential.

sds-overview.png

Explore this topic



Defining the need

The old ways of scaling are failing us.

Inverse hypervisor
We turn virtualization on its ear. Learn how it works.

Fast results and lower costs
And we’ll show you the benchmarks to back it up.

Simple administration
Point-and-click orchestration makes it easy.

More from the cloud
Learn how to get the most from cloud services.

Software-defined future
Servers are the missing piece of the Software-Defined Data Center.

Roll your own
How to build a Software-Defined Server.

Why Software-Defined Servers are needed now more than ever.

Data is growing—fast. More data was generated in the past two years than in the previous 5,000 combined. This has prompted data sets of staggering size and problems too big or complex for even the largest servers to handle.

 

Processor technology isn’t keeping up. “Moore’s Law,” which for 50 years held that microprocessor performance doubles every 18 months, is effectively dead. Despite multicore designs, parallelism and other heroic tricks to goose overall speed, processor performance improvements are leveling off at the very moment they need to skyrocket.

sds-rack

Software-defined data centers are lacking the most crucial resource of all. Organizations are trying to handle large or unpredictable workloads by running them on composable, software-defined resources. But while today’s software-defined data centers (SDDCs) benefit mightily from software-defined storage and networks, the SDDC’s most essential resource –the server—remains fixed and inflexible. (Well, until now.)

We turn virtualization on its ear. Learn how it works.

inverse-hypervisor

Everyone’s familiar with classic virtualization—extracting more flexibility from a single server by using a hypervisor like VMware to run multiple system images on that one server. 

But what if you want to get more flexibility and utilization from multiple servers? Classic virtualization is no help there, because the size of your VMware virtual machine is limited to the size of the physical machine it’s running on.

HyperKernel: the inverse hypervisor. This is where TidaScale’s inverse hypervisor technology comes in. TidalScale’s inverse hypervisor—the TidalScale HyperKernel—binds multiple physical computers into a single, coherent virtual system. Your virtual system (we call it a Software-Defined Server) allows you to run an unmodified guest operating system, with no changes to your existing applications required. All the resources associated with the aggregated systems – memory, cores, storage and I/O – are available to you and your application. You can configure them as one system or several. Learn more here:

WATCH TidalScale technical overview
WATCH TidalScale HyperKernel
READ Software-Defined Servers: A guide to adding the missing piece in the software-defined data center puzzle (featuring research from Gartner)

Resource migration and real-time machine learning. TidalScale’s Software-Defined Servers involve more than aggregating resources. The HyperKernel uses patented, real-time machine learning technology to optimize performance by automatically migrating memory pages and virtual CPUs to where the application needs them. Learn how it works:

WATCH The Machine Learning Layer
READ 3 secrets to right-sizing a server

Lower latencies. The HyperKernel also takes advantage of the lower latencies of Ethernet networks and the memory management hardware acceleration in the latest CPUs (Intel’s VT-x and VT-d) to algorithmically and automatically tune the performance of Software-Defined Servers. Find out what’s driving our ability to dynamically tune performance:

READ 7 Trends Driving the Industry to Software-Defined Servers

Serve both the business and IT.

It’s not like data sets are getting smaller.

Unfortunately data centers are comprised of fixed resources and hard-to-scale assets that make it tough to meet escalating workloads and fluctuating requirements. Users and business units don’t have time to wait for the next big system procurement or major application rewrite to have your code run across clusters.

With TidalScale Software-Defined Servers, the wait is over. By aggregating multiple commodity servers into one or more virtual systems, you can size your computer to fit the problem. And you can do it quickly, so you can pursue those revenue opportunities straight away.

  • Run your memory-intensive problems on a shared memory space totaling up to 64TB, and compute-intensive problems on hundreds of cores. Learn more
  • Increase throughput by an order of magnitude. Learn more
  • Cut the time required to validate models. Learn more
  • Build and deploy a new Software-Defined Server in 5 minutes. Learn more

gartner-paper-graphic


How can you extract more value from existing assets?
 
 It’s a big focus for DevOps and IT management. Containers offer one way to do this, though containers alone don’t address all needs. So they look to cloud infrastructure-as-a-service (Iaas) providers to access on-demand resources they don’t have in house. Yet even with IaaS offering, options are limited.TidalScale helps you lower the TCO of both on-premise and cloud data center assets. You use (and pay for) only what you need, when you need it. And you no longer have to buy or lease unnecessary cores just to get enough memory for your workload. What’s more, you can cut software licensing costs, which are typically tied to core count. With TidalScale, you don’t have to sacrifice economy for fast results.

  • Increase performance on critical memory-intensive applications like R. Learn more
  • Optimize DevOps by reconfiguring software-defined resources on demand. Learn more
  • Deploy the largest cloud systems imaginable, and pay by the hour. Learn more
  • Drive down the TCO of sweet spot servers. Learn more
  • Turn storage and compute into strategic weapons for data center management. Learn more

 

Point-and-click orchestration makes it easy.

simple-admin

It’s understandable to think aggregating the collective resources of multiple commodity servers into a single system would be difficult. But it’s not.

With TidalScale WaveRunner, our point-and-click orchestration software, TidalScale has taken something that should be complex and made it very, very simple. 

How simple? With WaveRunner, you can:

  • Choose from available systems to create a right-sized Software-Defined Server in under 5 minutes and just 3 clicks.
  • Visually monitor how your aggregated resources are being utilized.
  • Allocate resources in real time, as they’re needed.
  • Respond to changing workloads by creating a new Software-Defined Server on demand.
  • Simplify management of all the virtualized resources in your Software-Defined Data Center (SDDC), including compute, memory, storage and networks.

Learn more here:

WATCH WaveRunner in action 
READ How WaveRunner puts you in control of a better SDDC
READ Why Software-Defined Servers and storage are a match made in data center heaven
READ Software-Defined Servers: A guide to adding the missing piece in the software-defined data center puzzle (featuring research from Gartner)

How to get more from your cloud infrastructure.

Tapping a cloud infrastructure is a popular way to scale capacity without scaling capital costs. But even your virtualized public or private cloud can suffer from the same lack of flexibility that plagues classic on-premise data centers. 

In fact, Gartner recently predicted that by 2020, more than a third of large enterprises will “implement dynamic optimization technology for private and virtualized infrastructure.” Gartner calls dynamic optimization—enabling composable resources configured on demand—“the next battleground for the cloud.”*

more-from-the-cloud

With TidalScale, that battle is easy to win. Our Software-Defined Servers allow you to dynamically size servers in the cloud. No longer are you limited by the largest configuration available from your IaaS provider. Now you can create servers of virtually any size from commodity cloud servers, with no changes to your OS or applications. And with our Wave Runner orchestration software, you’ll have total control over your virtualized, composable cloud resources.

  • On-premise or hosted, you can optimize your cloud. Learn more.
  • Keep up with workloads with a truly composable infrastructure. Learn more.
  • Use containers to deliver an efficient private cloud. Learn more.
  • Build the cloud’s largest servers with TidalScale on Oracle Cloud Infrastructure. Learn more.
  • Did you know you can deploy TidalScale as a Service? Learn more.

*Gartner “Cool Vendors in Cloud Computing, 2017” by Gregor Petri, David Edward Ackerman, Stanley Zaffos, Sid Nag, Ed Anderson, Cathy Tornbohm, [May 25, 2017].

The software-defined data center has been missing its most critical piece. Until now.

build-a-sds

Software-defined data centers (SDDCs) have become increasingly popular as organizations seek to achieve results faster, lower their costs and streamline DevOps and IT administration.

 

But while software-defined storage and networks are by now commonplace, servers—arguably the most critical piece of the SDDC puzzle—have remained a fixed resource. They’ve prevented SDDCs from becoming wholly software-defined.

TidalScale changes all that. With our Software-Defined Servers, you can dynamically size servers to accommodate any workload or data set. Use commodity “sweet spot” servers and standard interconnects. No software changes required.

Read what this means for your plans to implement your own SDDC.

How to build a
Software-Defined Server.

In the time it takes to grab a cup of coffee, you can configure and boot up a new Software-Defined Server. Better yet, you can use hardware you already own, or even use servers from a cloud compute provider. Here’s how in four easy steps.

1. Select the servers you want to combine into a Software-Defined Server. 
TidalScale combines commodity servers and interconnects into a single virtual system that shares all the aggregate resources of those servers. Simply select the servers you want to combine, whether they’re “sweet spot” servers you have on hand currently, or are the systems available from one of our supported cloud infrastructure providers. Need to purchase new hardware? You’re in luck: With TidalScale, you can build massive Software-Defined Servers from simple, cost-effective, commodity servers.

Learn more here:

READ 4 ways to right-size a server right now
READ It’s now easy to build HPC systems in the cloud
READ TidalScale on Oracle Cloud Infrastructure
READ TidalScale on OrionVM Wholesale Cloud Platform

2. Combine them using TidalScale WaveRunner.
WaveRunner is our intuitive, point-and-click orchestration software—your control panel for configuring, booting and managing your Software-Defined Servers. With WaveRunner, you can visually choose the nodes you want to combine to achieve the right level of resources—memory, cores, storage and network—that your workload demands.

Get the details here:

WATCH Using WaveRunner to configure a Software-Defined Server
READ Software-Defined Server Planning Guide, featuring Gartner Research

roll-your-own-2

3. Monitor resource utilization.
Now that you’re in control of your own software-defined resources, use WaveRunner’s sparklines and visual indicators to keep track of system status.

4. Reconfigure on the spot.
When one task is done, or when your workload changes, tear down your Software-Defined Server and stand up a new one just as quickly as you configured the first.

More ways to learn:

READ Why you need a BFC (Big Flexible Computer)
READ Scaling the computer to the problem (IEEE cover feature)
WATCH TidalScale technical overview

gartner_cool_vendor_2017.jpg
IDC-Innovator-Logo-2017.png
TidalScale featured in e-week!
rhglobal100logo_360.png
Gary Smeardon in the Cube Interview