While the IC’s research organization looks into adding security to cloud environments, in the here and now, intelligence agencies are sharing more data.
It’s one of the most important decisions facing today’s data center chiefs: Should our IT team move specific workloads off-premises or keep everything here in our data center?
The government’s Data Center Optimization Initiative, which went into effect last summer, is designed to save money and boost efficiency, but is also expected spur the adoption of cloud services and interagency shared services.
The new mandate requires agencies to “continue to principally reduce application, system, and database inventories to essential enterprise levels by increasing the use of virtualization to enable pooling of storage, network and computer resources, and dynamic allocation on-demand.”
After that, agencies are required to explore options for consolidating or closing existing data centers by transitioning to provisioned Software as a Service, Platform as a Service and Infrastructure as a Service models “to the furthest extent practicable, consistent with the Cloud First policy.”
As server virtualization continues to make it easier to move workloads to the cloud, organizations are now migrating elements of IT to the cloud to reduce their data center footprint, notes 451 Research Director Peter Christy in a recent survey report from Uptime Institute.
Yet as more agencies investigate cloud environments with an eye toward boosting efficiency and cutting costs, many agency IT leaders are also discovering that making the right choices can be complex. Today’s market offers multiple options for where and how to run workloads. The reality is that certain workloads are better suited to on-premises environments while others can run more productively in the cloud.
“It’s an architectural shift,” says Bobby Patrick, chief marketing officer for HPE Cloud at Hewlett Packard Enterprise. “Whether on- or off-premises, you’re trying to create an IT environment with a significantly lower operating cost structure and significantly higher development and power. If you don’t accomplish that, then you’re not successful, regardless of what technology you use.”
The options weren’t always so extensive. Until about a decade ago, on-premises solutions were really the only way to go. In an on-premises-only world, servers were acquired, operating systems installed and various types of related hardware and software (such as storage subsystems and network gear) completed the picture.
In that former technology era, just about everything that mattered was contained within the data center’s four walls. IT staff members could easily reach out, touch and service, or modify all equipment and solutions. An agency’s leaders could proudly show off their flashy data centers to leaders from other agencies, potential employees and other stakeholders.
Even after servers began to be virtualized, the physical hosts remained. An agency might decide to place routine data center management responsibilities into the hands of an outside party, but the agency itself was still in charge of the facility’s physical security, power and cooling resources, and basic infrastructure. When some agencies began moving their hardware into cages located within shared data centers or hosting facilities, they still retained responsibility for their tangible, physical assets.
Today, it has become relatively easy to move workloads to cloud providers that guarantee acceptable performance at attractive price points.
“With the cloud, there are a lot of benefits — the primary one being the ability to scale the service up and down as needed quickly, assuming the application can be scaled dynamically,” says Mike Fratto, principal analyst for enterprise network systems at Current Analysis.
Fratto notes that agencies and other organizations “can get more cost-effective uptime and a lower time to recovery if a failure occurs, assuming the service has adequate features and the customer takes advantage of them.”
All enterprises can benefit in some way by moving workloads to the cloud, adds Lori MacVittie, principal technical expert at F5 Networks. “Whether it’s sending commoditized business operations to Software as a Service or a small business putting its web presence in the cloud or taking advantage of Office 365 and storing files in the cloud, there are definitely workloads that are more cost-effective in the cloud than on-premises in every organization.”
Over time, the cloud has split into two distinct models: public and private. The public cloud made life easier by placing the cloud provider in charge of most essential operations, such as management and maintenance. A number of agencies have also been drawn into the public cloud by the fact that they can use the approach to reduce lead times in certain areas, such as testing and deployment of new products.
Nevertheless, many agencies continue to view public-cloud providers with skepticism, believing that the model introduces new intrusion, privacy, control and compliance challenges (despite the fact that security breaches of public clouds have been relatively rare).
The on-premises private cloud model has dispelled most of these misgivings, because it can reside on an agency’s own intranet or at a hosted data center behind a firewall. Unlike multitenant public clouds, which deliver services to many agencies, a private cloud is dedicated to a single organization.
“Organizations can build a private cloud that is every bit as robust and capable as a public-cloud service,” Fratto says.
Private clouds appeal to many organizations that already have fully built-out data centers because they can continue using their existing infrastructure. But that private cloud also means the IT team must continue to handle management responsibilities, maintenance and systems updates — all of which cost time and money. Some other expenses to consider are necessary outlays for virtualization, cloud software and cloud management tools.
“One way to look at on-premises versus off-premises is like this: Public clouds, off-premises, give a lot of convenience, but you sacrifice a lot of control. Private clouds give a lot of control, but you don’t quite have as much convenience,” Patrick says.