Despite what many organizations have assumed, the public cloud is not the best option for all workloads and use cases.
On-premise object storage continues to gain ground as organizations across a variety of industries turn to the technology to accommodate their long-term storage needs. While hyperscale public cloud storage services may appear attractive, many federal, state and local government agencies now find that on-prem object storage provides important benefits in cost, performance, security and control. Because those considerations are all critical for government entities, we’ll take a closer look at how on-prem object storage compares to the public cloud in each area.
Cost considerations are paramount for government agencies, which have seen their IT budgets tighten over the past couple years. Many organizations have learned the hard way that the public cloud is more expensive than they anticipated when they suddenly start receiving larger-than-expected monthly bills. The problem is that using the public cloud comes with many hidden costs.
Beyond the storage fees, there are data egress fees and data replication fees. These tend to average around 15-20% of an organization’s cloud bill but can grow higher in storage-intensive workloads. This is in addition to the price of the WAN connection to the cloud. Network bandwidth and data access charges grow particularly high when performing computing tasks on large data sets, such as running artificial intelligence and machine learning workloads or even doing simpler analytics operations on a large database. The more data a user accesses, the higher the bill.
For storage usage that exceeds a few hundred terabytes of capacity, on-prem object storage is often a better option than the public cloud because it doesn’t entail these expensive network bandwidth, data access and data egress fees. There are no special, hidden costs to connect to data or run capacity-intensive workloads for critical use cases such as artificial intelligence and machine learning. In addition, object storage costs are also far more predictable, which is an important factor for government agencies with rigid budget constraints.
Like cost, performance in the public cloud can impact overall application performance. Public cloud performance, which can be broadly described as the time it takes to transfer data to and from the cloud, is dependent on the available WAN bandwidth and how many total workloads are being run on a cloud provider at a given time. Capacity-intensive workloads in particular will experience high latency due to the time required to move data to and from the cloud.
Here’s a way to estimate what this means: To move 1 TB of data over a 1 gigabit link takes about 180 minutes, and these capacity and bandwidth figures can be adjusted to match different use cases. When agencies consider their application requirements, they should be sure if they’re thinking about backing up data to the cloud, for example, that they can meet their backup performance objectives with the recovery time they’ve calculated.
On-prem object storage provides local access to data over multiple high-speed network connections that operate in parallel. And because on-prem object storage is limitlessly scalable, governments can easily deploy the capacity they need to run data-intensive workloads at low latency.
Security and control
Security and control are just as critical for government agencies as cost and performance. When it comes to the public cloud, there’s a great deal of confusion regarding who’s responsible for managing security. Many public cloud users think their cloud provider ensures that their data is kept safe. However, this isn’t actually the case, and this misunderstanding has left many organizations vulnerable.
Furthermore, best practices for securing public cloud deployments are very different than best practices for securing traditional on-prem deployments, which organizations have grown accustomed to over decades. The learning curve makes many new public cloud customers susceptible to breaches as they work to get up to speed. To cite a common example, these new customers often will incorrectly configure cloud instances within cloud storage buckets, leaving critical data exposed.
All of this highlights the need for government agencies to take charge of their own security to protect their data, and the best way to do so is keeping it in an on-prem data center or private cloud, where they have full autonomy over all security and data access policies.
Object storage enables best of both worlds
The public cloud provides significant agility and convenience benefits. It’s a great fit for certain use cases, such as disaster recovery, which can be done cost effectively in the public cloud. However, despite what many organizations assumed in years past, the public cloud is not the best option for all workloads and industries. This is particularly true for government agencies, which must prioritize cost, performance and control of their data.
That said, because most on-prem object storage platforms are built around the S3 API – the de facto language of public cloud storage – they’re highly compatible with the public cloud. As a result, governments and other organizations can embrace a hybrid approach by starting with object storage on-premises or in a private cloud, then expanding those deployments to the public cloud to support certain workloads. Of course, it’s a two-way street: while on-prem object storage deployments can seamlessly grow to the public cloud, deployments that started in the public cloud can also be easily migrated to on-prem object storage.