Containers: The Case for VM Coexistence
Today, corporations have the technical flexibility to customize the datacenter to better fit their long- and short-term business goals. They are no longer locked into legacy infrastructure upgrades as the singular or even the primary source of technical improvements. Adopting new technologies and new movements in computing styles are less likely to involve long waits for vendor advances in basic infrastructure. Fundamental improvements can occur in middleware, the software that manages the delivery of data services to users and customers.
The growth of virtualization and related emergence of private/public/hybrid clouds have transformed the database management landscape. Rather than the slow-moving infrastructure play it once was, the database is changing to more efficiently support the on- and off-premise stakeholders generate.
VMware is the unquestioned leader in the hypervisor/virtual machine (VM) market. However, while it is currently the centerpiece of the VM market, as database expectations grow its grip as the singular linchpin of virtualization is loosening.
Enter containers – the ideal complement, even lightweight alternative – to traditional VMs in a computing world where agility is at a premium.
VMs and the Cloud Changed Everything
There are multiple stakeholders in the business of corporate database management. On the administration side there are business managers, departmental managers, and IT managers. On the other side, there are users, and most importantly, customers.
Not long ago database management came in a single flavor. It was hardwired, server-based, and IT had only so much flexibility in determining the quality of the product being delivered.
The emergence of VMs and the cloud changed everything. They added a layer of datacenter flexibility. Service delivery was no longer based entirely on the ownership of infrastructure. Now, a departmental manager or entrepreneur with a credit card could set up a cloud-based database management system in minutes without touching a server or even talking to IT.
That flexibility raised expectations among database stakeholders, particularly customers. They have grown accustomed to the ease of use VMs brought to database access. Easy data access empowers users and raises corporate managers’ expectations – driving personal and corporate creativity and innovation.
The Case for Containers
There is no single technology or vendor that fully governs the variety of changes currently occurring in the modern datacenter. New demands on the corporate database are emerging from a number of divergent sources. For instance, the concept of the Internet of Things (IoT) is becoming more real. It is ushering in an era of new applications that will employ the datacenter as a critical resource for things like real-time data from billions of sensors and other embedded systems.
Unstructured data, including images like x-rays and MRIs, force drastic changes in databases.
Also, once limited to structured data like accounting reports, the datacenter is being forced to expand to include unstructured data like MRIs, sound files, and video. That movement is driven in large part by the medical industry’s ongoing transition from paper to digital records. This rich data avalanche – highly dependent on images, interactivity, and new datasets -- is a massive challenge for datacenter administrators.
These changes have not been easily absorbed into the existing datacenter infrastructure. As a result, we have seen the emergence of technologies like Apache Hadoop which is designed to process large unstructured datasets. Hadoop does not compete directly with the relational database, but it more efficiently processes and analyzes data-types that are foreign to relational technology.
Traditional virtualization and containers present a similar technical conundrum/opportunity. Like the relational database, VMs were designed for a specific set of circumstances and continue to operate within a certain range of speed and flexibility. But there is growing demand for a new mode of operation in the datacenter: one that is faster, more flexible, and much more responsive to the day-to-day competitive realities businesses face. CIOs know this state of affairs well; they get the complaints.
Bimodal IT
Research firm Gartner defined a solution, and now numerous other industry pundits are promoting its adoption: Bimodal IT.
Bimodal IT divides IT into two parts. Mode1 zeroes in on IT’s traditional strengths: maintaining and improving enterprise applications, infrastructure, and services. That includes functions like network scalability, safety, efficiency, and accuracy. Mode2 focuses on the ad-hoc opportunities of the digital economy: the I-need-it-yesterday business pace of the Internet. The endgame for Mode2 is agility, speed, exploration, and innovation.
(Source: Gartner)
A main indicator of the need for this two-pronged approach is the fact departmental managers continue to complain about too-slow datacenter response times or in more extreme circumstances they bypass IT when new projects come up.
Containers - Bimodal Bridges
Containers are infrastructure-agnostic so they offer customers a solution: a bimodal bridge.
In Mode1, containers can be used in traditional settings as either lightweight VM replacements or enhancements. They can significantly reduce the number of VMs and operating system instances IT has to manage by containerizing and stacking workloads across fewer, larger VMs.
In Mode2, containers offer the light footprint that affords IT added agility and flexibility across new systems and platforms. In applications involving rapidly-emerging data streams like electronic medical records, containers can make data more highly available, mobile, and less costly to deploy and distribute.
VMs weren’t built for smooth operation in Mode2. They were designed to improve server utilization by encapsulating each application in a VM to permit multiple virtualized apps to run on the same physical hardware. One has to provision infrastructure elements like networking and storage in order to host the application. These requirements increase the size of the VMs. Containers virtualize the application and not the machine, storage, or networking, so they are lighter than traditional VMs. This relative lightness allows for higher application mobility across virtually any infrastructure. This light footprint also allows for faster backup and restore and consumes less memory.
IT can more efficiently customize data services on the fly – this, at a time when CIOs are focusing on datacenter agility and innovation. They also seek increased scale to keep up with the rapidly growing amount of data businesses are churning out. Containers help CIOs simplify maintenance and quickly free-up unused or underutilized resources for new initiatives.
Using containers as a bridge, Mode1 can act as either the engine for traditional data requests, applications, and operations or as a transparent back-office for Mode2 operations. Mode2 users frequently need access to traditional business applications and stored data.
As a bimodal bridge, containers are uniquely positioned to enable these two distinct IT methodologies to succeed in the same organization – many times across two or more separate teams. The traditional IT team maintains management of day-to-day functions, while the agile data team supports the business lines’ blossoming requirements. Traditional IT continues to develop long-term strategies and ensures successful technology acquisition, deployment, and management. The agile data team guarantees business lines are kept constantly ready and highly adaptable to the ever-changing technology landscape and consequently, prepared to respond, innovate, create competitive advantage, and capitalize on emerging business opportunities.
About the Author:
Don Boxley is CEO and co-founder of DH2i. Prior to founding DH2i, he spent more than four years at Hewlett Packard, and founded another start-up, DCBox Marketing Services. Don specializes in: P&L management, executive team building, cross-team communication, sales strategy and execution, demand generation, marketing strategy development, product requirement definition, product management, messaging, field marketing, business development, and channel development. Follow the company on Twitter @dh2i.
