Cloud infrastructure providers today don’t have much flexibility when it comes to the systems they use. Resources devoted to running specific applications and workloads are generally confined to the limits of a single system, typically a “sweet spot” server that perhaps offers 24 cores and a few hundred gigs of memory. It's a matter of economics, really: If the vast majority of your hosted workloads can be handled by a sweet spot server, it makes little financial sense to invest in large HPC-class servers. After all, they didn’t give it the name "sweet spot" for nothing.
But as we all know, data sets are growing, and Big Data analytics are becoming an essential part of doing business in our data-driven era. That means the resourcing strategies in use today will probably fail to provide the flexibility needed to meet the fluctuating demands on data centers and hosted infrastructures.
For infrastructure providers to achieve the kind of agility they need to accommodate spikes in workloads and escalating application requirements for more cores and memory, they need a way to push past the limits of typical server configurations whenever the need arises. They need to configure large, even huge servers on demand – and then run their applications on those large servers without recoding, sharding or other productivity-killing techniques.
TidalScale has figured this out, and its Software-Defined Server platform delivers exactly what infrastructure providers need to do more than ever before. If you’re not familiar, a Software-Defined Server combines a number of industry-standard, commodity hardware servers – let’s call them “sweet spot” servers – and binds them together into a single, large virtual system. I like to think of this concept as inverse virtualization, essentially using a hypervisor to aggregate multiple physical machines rather than chopping them up.
In TidalScale’s case, this is accomplished without requiring a single change to your application or OS. All resources associated with the combined servers, including cores, memory and I/O, are available to the application, which remains totally unaware that it is actually running on four, eight, even dozens of different servers. TidalScale’s self-optimizing platform uses advanced machine learning to migrate resources dynamically as workloads require. Users can configure and boot a new Software-Defined Server in minutes, and reconfigure it just as quickly. And if your needs require you to boot multiple servers of varying sizes from your available pool of individual servers, TidalScale enables that, too.
All this is achieved without having to step into the realm of non-commodity infrastructure – good news for providers trying to keep operating costs low.
3 Key Benefits to Infrastructure Providers
My company, OrionVM, provides infrastructure-as-a-service (IaaS), so whenever I learn about a new breakthrough that brings new levels of agility and ease of deployment to my data center, I think about the larger impact on infrastructure providers as a whole. And for them, I believe Software-Defined Servers are a game changer because they deliver an entirely new level of flexibility in three key areas:
- No more ‘one size fits all.’ Aggregating several standard systems into one allows infrastructure providers to offer greater flexibility to their customers. Now you won’t be stuck with the “one size fits all” solution because you need to limit your number of product lines in inventory.
- More SKUs, but without adding more servers. You have the freedom to offer more while using what you already have. Rather than keeping dedicated pools of small, medium and large servers, any number of these machines can be aggregated to deliver a much larger number of SKUs. This relieves the inventory problem that so many dedicated inventory providers face.
- You can use standard hardware. TidalScale designed its Software-Defined Server platform, known as HyperKernel, to leverage industry-standard components (e.g. x86 machines and 10 Gb Ethernet interconnects); in other words, just the standard technologies that make up the sweet spot servers that populate every data center. This keeps infrastructure costs low and makes it easy to repurpose these machines across a variety of workloads.
Taking Software-Defined Servers to the Cloud
Upon recognizing the vast benefits of Software-Defined Servers, OrionVM has integrated the TidalScale solution into the OrionVM Wholesale Cloud, allowing TidalScale customers to leverage the flexibility of cloud infrastructure with the scale and scalability of a TidalPod – TidalScale’s name for a deployed Software-Defined Server.
OrionVM’s platform provides ultimate flexibility to both OrionVM and TidalScale customers, allowing them to leverage the combined benefits of our offerings, including:
The ability to access Software-Defined Servers ranging in capacity from 1.3TB to 13TB of DRAM and up to 400 cores – instances that are far larger than any available from ordinary cloud vendors, yet they are available at lower cost per gigabyte of memory.
- Persistent, high-speed and fully redundant storage volumes ranging in performance from slow archival storage, through to high-speed SSD storage.
- Ability to tightly interconnect smaller Cloud Instances with large scale TidalScale VMs all on one network to allow high-speed data access with low latency and cost.
- Flexibility to mix and match storage performance tiers on the same Cloud Instance or TidalScale VM.
- Ability to seamlessly attach storage volumes to Cloud VMs to perform data ingest, and then easily present these storage volumes to a TidalScale VM.
- Ability to easily ingest data via high-speed internet links, via cross connects to various cloud providers (e.g. AWS/Azure/GCE), or via a private link to your existing datacenter environment.
The partnership between TidalScale and OrionVM revolutionizes the type of applications that can be hosted on the cloud, allowing customers to scale as required by their application – all via a single portal.
As I say, for infrastructure providers, this is a real game changer. The sweet spot just got a lot sweeter.