2015: Cloud automation turns from operations to applications
- By Carolyn Duffy Marsan
- Dec 22, 2014
As cloud deployments gain momentum across government, agencies are benefitting from increased levels of automation that make it easier, cheaper and faster to spin up new IT resources and deploy cutting-edge web applications.
In 2015, more capabilities will be built into government cloud offerings, such as support for security and privacy standards as well as for hybrid public/private cloud deployments.
What’s more, the development of orchestration layers to improve data integration between cloud-based applications and the push towards network virtualization, will pick up speed.
“In the last year, public clouds have emerged that can better support compliance-driven organizations,’’ said Bill Kleyman, national director of strategy and innovation at MTM Technologies, a Stamford, Conn.-based consultancy.
“Cloud providers are now offering solutions that support FedRAMP (and the cloud-first initiative) so agencies can utilize all of this really great infrastructure specifically built for government cloud specifications and use cases.’’
A case in point: Microsoft announced in December that its Azure Cloud for Government supports such standards as FedRAMP, FISMA, DOD Enterprise Cloud Service Broker, HIPAA, IRS 1075 and CJIS.
Microsoft and Amazon Web Services (AWS) also offer tools that make it easier for agencies to create hybrid systems, where some applications reside in government-run clouds while others are in public clouds.
Amazon’s autoscaling feature, for example, makes it easy for agencies to handle peak usage in the public cloud as needed – at less cost than dedicated IT resources.
“A government organization can go to a private cloud with an application and say that when this reaches an 80 percent threshold, automatically provision 4G of RAM to support it,’’ Kleyman said.
“In the past, that application was sitting on a server, and it sent an alert to administrator that it had just hit a threshold. Now the administrator doesn’t have to sit down and do everything manually.’’
And improved autoscaling is due soon. Amazon is previewing a capability called AWS Lambda, which starts running milliseconds after an event such as a website click and automatically triggers compute resources.
The service is designed as a cost-effective way for a web app to scale from receiving a few requests a day to thousands per second. It could be useful when law enforcement agencies interact with citizens, such as asking for videos after the Boston Marathon bombing.
New government workloads to the cloud
Agencies are also migrating new workloads to government-specific cloud offerings. Initially, agencies used cloud-based services for development/testing, disaster recovery and bursty applications like video storage. Now data analytics and web applications are migrating to the cloud too.
“More government agencies are going to be creating apps because they are the easiest way to get to their end users: the taxpayers,’’ Kleyman said. “All those apps are going to generate more data, and those apps are going to be cloud-hosted because of resource constraints.’’
For example, the National Institutes of Health’s National Database for Autism Research (NDAR) built a cloud-based collaboration platform using AWS to replace an outmoded system of mailing copies of data stored on hard disks. Researchers now access data through AWS, which automatically stands up a processing environment and provides analytic tools.
“One of the main benefits is that the NIH has more security,’’ said Mark Ryland, chief solutions architect for AWS’s worldwide public sector team. “They know who is accessing data now, and they can shut them down if they need to because some of this data is very sensitive.’’
The cloud-based approach also means more researchers can collaborate on the NDAR database.
“The bottom-line benefit is much faster time-to-science,’’ Ryland said. “There will be more collaborators because it is easier for smaller and medium-sized universities to get involved. They can do this for $20 a day for infrastructure versus creating a physical- or capital-intensive infrastructure.’’
With cloud-based access to its data, NDAR is leading a culture change within the NIH toward increased data sharing, says Dr. Tom Insel, Director of the National Institute of Mental Health.
“Virtually all autism human subjects research data is expected to be deposited in the National Database for Autism Research, which now holds genomic sequences, brain images and clinical data from over 77,000 subjects,’’ Insel said in a recent blog post. “This data provides a platform for discovery through secondary analysis and data sharing specific to a publication.’’
Next: application level intelligence
In addition to government-specific cloud offerings from Microsoft and AWS, the market is packed with cloud management platforms from CSC, RightScale, Cisco, IBM, VMware and others. These tools add an orchestration layer that allows agencies to manage the cost and use of cloud-based IT assets.
The return on these tools is improved efficiency, said David Linthicum, senior vice president of Cloud Technology Partners, a Boston consultancy. “You have an orchestration or automation layer managing these assets versus a person sitting down at a console spinning up cloud or non-cloud resources,’’ he explained.
Agencies have been using these tools to automate the provisioning of compute, storage and networking, including firewalls and load balancers. But the deployment of application-level intelligence is on the horizon for 2015, Linthicum said.
“The first step is provisioning and deprovisioning. Next is adding application-level intelligence into the process,’’ he said.
“Imagine each application is a little silo that does specific things with a static process and data bound to it. A CRM app like Salesforce, a battlefield management app and an HR system are all silos. But what if the apps had the ability to leverage each other’s processes, behavior and data? Building a meta app like that is possible with some of the orchestration systems.’’
Bill Rowan, vice president of federal at VMware Public Sector, said the biggest boon for agencies with these tools is in the automation of network provisioning. Previously, organizations hard-wired compute and storage resources to a particular application through a patch panel. Now these configurations can be changed on the fly through network virtualization.
“The biggest bang in terms of changing the way agencies operate is automating the network process – where the people and process time is spent is on the network,’’ Rowan said. “A year from now, I think we will be surprised at how many customers … have moved to the automation of network provisioning.’’
Carolyn Duffy Marsan is a writer based in Milwaukee, Wisc., covering enterprise technology.