How can the New York City Department of Youth and Community Development (DYCD) use information and technology to achieve its mission of alleviating the effects of poverty and providing opportunities for New Yorkers and communities to flourish? As Associate Commissioner of Information Technology at DYCD, I focus on this question every day.
There is a delicate balance that must be maintained between spending the agency’s budget wisely while still providing high-quality technology solutions that can improve the lives of New Yorkers. Decisions are constantly made to determine where best to invest our technology dollars, whether it is on internal resources or through service providers, namely via the Cloud. There are several factors to be considered when deciding between using on-premise tools or theCloud:
Ease of Access/Performance
Users expect their eGovernment services to be as user-friendly and perform as well as the commercial products that they use on a regular basis. They want to use a variety of devices and platforms, including mobile, to access resources as quickly and easily as possible.
Having an application in the Cloud provides the technology to easily create publicly available applications that are accessible from anywhere. It also offers the ability to scale up or down with ease, assuming the application is built for the Cloud.
Using a well-known cloud service provider enhances your overall security. It alleviates the need to secure the utility aspects of IT, starting with hardware for infrastructure services, through OS and even application layers for Software as a Service. This allows the customer to focus on securing the most important aspect of their application: the data.
Large cloud providers have focused teams reviewing security every day—more than a small government agency can ever hope to accomplish. Though be wary: not every cloud provider has the same level of security for their systems.
System uptime is extremely important when delivering applications. If an application is frequently down, the public will not use the resource. For internal applications, downtime causes productivity loss and missed deadlines.
Availability needs to also include business continuity and disaster recovery. Loss of company-owned hardware is much more difficult to replace than a cloud-provided virtual machine. For disaster recovery, data can be saved on the Cloud across the world with a click of a button. You no longer need to ship data on a tape to an offsite facility right before a hurricane, only to spend another six days to recover from the tape if data were lost.
“From what I have seen, you will likely need both Cloud services and an on-premise datacenter, even if scaled back, to deliver your services”
Total Cost of Ownership
There are many models available to calculate Total Cost of Ownership (TCO), but it is always hard to find those hidden costs.
When I think about TCO between the Cloud and on-premise, I focus on not just licensing but the personnel to procure and maintain facilities and security operations.
• Procurement: This includes hardware and software licenses, and personnel cycles from not just IT but Procurement, Administrative Services and Fiscal. From creating just one contract to do everything you need with a cloud provider to possibly 10 contracts, deliveries, etc. for the same service if it were built internally.
• Facility: Beyond just space, increasing hardware usage will increase costs for power, cooling, site maintenance, temperature monitoring, battery backup, backup sites for disaster recovery, and the resources to be available in an emergency (i.e. overtime costs, vendor contracts, etc.), to name a few. Some of this can be offset with converged and hyper-converged hardware which will reduce the needed footprint.
• Resources: Maintaining a datacenter requires resources that can handle multiple technologies including hardware, firewalls, switching, routers, backup, disaster recovery, etc. While some resources can wear multiple hats, with every new hat they wear and every new technology that is introduced, it reduces their overall effectiveness and/or requires additional personnel.
• Security: To constantly monitor and update for security vulnerabilities alone requires a team. Add in the tooling, after-hours support, tracking and remediation, and training, the costs continue to rise. Using a reputable cloud provider can dramatically reduce the vectors needed to monitor, provide much of the tooling, and allow you to focus on securing the areas that are most important.
Cloud services have been around for some time and most of my points above are widely known. Cloud providers have adapted to the needs of their customers, and the Cloud is more and more an extension of a company’s physical datacenter, providing the benefits of the Cloud while still being able to maintain workloads internally when needed. From what I have seen, you will likely need both Cloud services and an on-premise datacenter, even if scaled back, to deliver your services. Newer development methodologies such as DevOps need the flexibility to adjust and move quickly, which most internal datacenters cannot provide.
The Cloud is not for every workload, however, and moving to the Cloud with existing workloads can be a big lift. When considering whether to build out an on-premise datacenter or use a cloud provider, don’t forget the hidden costs that go beyond licensing. Assume that cloud costs will increase the more you consume but so will on-premise datacenters, where cost benefits will only be realized when you hit economies of scale.
You must always weigh your options carefully, but the Cloud is here to stay and should always be put into the mix when making technology decisions.