The folks from Springs.IO got in touch with me recently to discuss a new containers-based infrastructure as a service (IaaS) offering that they’re styling as “Reactive Servers.” The goal is offering small to medium sized businesses infrastructure services without requiring that they pay for more than they actually used or have expertise on staff to provision and manage a complex virtual environment.
Here’s what Springs.IO had to say
“Springs.io, a new simple, container-based cloud infrastructure service with usage-based billing, was launched today. As the only pure-play pay-as-you-use solution in the market, Springs.io will be targeted at Linux developers, agencies, and SMEs looking for low-cost, auto-scaling compute infrastructure.
Springs.io builds upon ElasticHosts’ next generation auto-scaling container technology, which allows customers to elastically expand capacity as needed, without the need for added software or configuration. Traditionally, cloud servers have been paid for according to the levels of capacity provisioned; if you run a 4GB server, then you pay for a 4GB server, regardless of whether that capacity is being used. Yet the container technology used by Springs.io enables a new model of operation, with billing to be based on actual usage, rather than provisioned capacity. For example, you might just pay for 1GB when a 4GB server is lightly used. These containers boot up in 2 seconds and offer bare metal performance without the overhead of virtualisation. Springs.io uses namespaces and cgroups, the same Linux kernel containerisation technologies as Docker or LXC, backed by high I/O SSD storage throughout. As a result, Springs.io can allow auto-scaling, high performance Linux containers at a fraction of the cost of traditional virtualization-based cloud servers.”
Springs.IO appears to recognize that virtual machine software isn’t the best answer for the creation of every virtual environment. It also appears to understand that its target audience, SMBs, are unlikely to have either the need or the expertise to make the best, most cost effective, use of virtual machine-based virtual environments.
As I’ve discussed in the past, containers, such as those offered by Dockers or Rocket, are implementations of a type of processing virtualization called operating system virtualization and partitioning. Unlike virtual machine software, this type of virtualization relies on the services of a single operating system. Each workload is isolated and set up as a separate partition under the control of a single operating system. In the past, this type of technology as been called LPARs, VPARs, or containers. This type of technology has appeared in Mainframe operating systems and UNIX for decades. It is also available in Linux and Windows.
As I mentioned in my article The Death of Virtual Machines? Not so Fast published in my column on VirtualizationReview.com, it serves a similar, but different role in the enterprise data center. Here’s a segment of what I had to say there:
The benefits of this approach are performance and efficiency. Because one OS is supporting all these virtual environments, less memory and storage is required. Furthermore, switching from one partition to another is very fast. Also, a single physical host may be able to support several times the number of independent workloads than if VM software was being used.
While using this approach, all of the workloads work with the same OS and the same version of that OS. So mixtures of Windows, Unix and Linux workloads on the same physical host aren’t part of the repertoire of this type of processing virtualization. The tradeoff is between performance and flexibility.
Springs.IO appears to want to make this approach easily available to enterprises that don’t want to deal with the overhead and complexity of using virtual machine software.
After reading through the company’s release I had a few questions. Here are the questions and Springs.IO’s responses:
After reading this, it occurred to me that Google Engine already supports containers. How is this the first?
Google container engine, i.e. https://cloud.google.com/container-engine/ is a service where they run Docker containers on top of standard virtual machines. This means you have the overhead of the virtualisation *and* the Docker engine. What it doesn’t do is take a Docker-like Linux container and put an entire operating system directly inside it, then have that run on the host with almost no overhead between it and the bare metal. That is what Springs.io does. Springs.io offers Operating System Containers, not Application Containers like Docker or Rocket. This is the same underlying technology, but instead of applying it to application portability and deployment, we have applied it to cloud hosting and virtualisation. There may be other hosting providers who have done something similar to Springs.io but I am certain that none of them offer a usage-only payment model where you only pay for the resources actually used, not the maximum provisioned capacity of the server.
Another point is that SMB customers often use packaged applications that aren’t developed to support containers. How is this service of interest to them?
Springs.io doesn’t use application containers, rather it uses the same underlying technology that powers Docker (cgroups and namespaces) as a way to get better performance, security and cost-effectiveness out of cloud servers. It features auto-scaling, which is of particular interest to SMB customers since with Springs.io there is no need to set up load-balancers, API scripts to scale horizontally, decide on what size your server should be, agonise over getting the right server package or setting the disk size correctly – instead you are billed no more and no less than the resources your server actually consumes. If you have a slow day and your server undergoes light load, you pay less. If you have a busy day and you’re getting a ton of traffic, you don’t need to worry about your server setup – it just scales up automatically and you pay for the extra resources for as long as they’re being used, then it scales back down.
I’d suggest that SMB IT decision-makers drop by the Springs.IO website to learn more. It appears that this service has the strong potential of making life easier for them.