In this special guest feature, Venkat Venkataramani, Co-founder and CEO for Rockset, explores the benefits of serverless with a focus on the top pain points that serverless technology eliminates, so that developers can focus on creativity and how they can improve their apps. Venkat was previously an Engineering Director in the Facebook infrastructure team responsible for all online data services that stored and served Facebook user data. Collectively, these systems worked across 5 geographies and and served more than 5 billion queries a second. Prior to Facebook, Venkat worked on the Oracle Database.
One of the biggest shifts in the IT industry has been businesses adopting public cloud infrastructure. The IDC forecasts that annual spending on public cloud infrastructure and services will achieve a five-year CAGR of 22.5% reaching a market size of $370B by 2022. Businesses move to the public cloud to increase agility and reduce operating costs, but these wins are contingent on the cloud adoption strategy. The harsh reality is that forklifting on-premises IT stacks and placing them on virtual servers in the cloud does not make businesses more agile, data-driven, or operationally efficient. Meanwhile, modern companies have quietly embraced a whole new world of operational analytics, which completely transforms the way data is collected and consumed by the business.
Not-So-Good Old Days
Procuring new hardware for data centers typically took months of planning as companies struggled to forecast their needs. Under-provisioned systems failed due to reliability issues while over-provisioned systems punched a hole in the IT budget. It was a constant struggle to manage performance while keeping costs under control. Public cloud infrastructure offered a cure by spinning up virtual servers on demand, thereby eliminating the need for long planning cycles. But in effect, public clouds solved one problem by creating a bigger problem. When virtual server instances became so easy to spin up, it led to massive server sprawl. This means IT teams needed to administer and secure a lot more virtual servers. The challenge of administering virtual server instances grows exponentially with each new cluster that a team needs to manage.
Lift, Shift and Get Left Behind
The challenge of managing virtual servers is compounded by the fact that the traditional IT software stack built for on-premises data centers is simply not designed to exploit the hardware fluidity of the cloud. Fork-lifting an existing software stack from the data center to dedicated virtual servers in the cloud is akin to living in an AirBnb full-time for three years or renting a car everyday for your 100-mile commute to work. Moving to the cloud using the “lift and shift” approach results in heavy overhead costs and leaves companies questioning their move to the cloud.
Simply put, if your cloud software stack does not take advantage of inherent hardware fluidity available in the cloud, then you are leaving a lot on the table.
Open Source Software on Cloud Hardware is the New Legacy
Companies migrated to the cloud to reduce operational costs and increase agility—a future that is possible with a serverless cloud architecture that transparently exploits cloud hardware fluidity and turns infrastructure into a utility. Serverless data infrastructure frees companies from server management and increases agility by democratizing access to operational data across the organization.
For example, embracing modern cloud data stack means moving away from managing your own HDFS cluster in the cloud and instead embracing serverless data offerings like Amazon S3 and Athena. Installing and managing your own open source software cluster on cloud hardware is the new legacy.
Go Serverless to Win in the Cloud
The winning cloud adoption strategy is to embrace not just cloud hardware but entire cloud services – so that your infrastructure stack functions reliably without you having to manage virtual servers. The term “serverless” continues to be used primarily to describe function-as-a-service compute frameworks such as AWS Lambda et al. But the narrative is changing, with modern data stacks going completely serverless as well. When a company installs and configures an open source database cluster in the cloud, it also needs to invest in continuous capacity planning and database tuning to constantly manage cost and performance. Contrast this to modern, fully managed cloud data services where there are no clusters to manage and performance can be taken for granted, such as cloud object stores and serverless search and analytics engines.
The benefits of serverless data are numerous and proven:
- No provisioning – Users shouldn’t have to concern themselves with what type of hardware they need to provision to set up the data management system.
- No capacity planning – Users shouldn’t need to plan cluster capacity at any point during the lifetime of the application. This means situations such as over-provisioned capacity burning a hole in their pockets or under-provisioned capacity causing performance and reliability issues should not be possible.
- No scaling limits – Users shouldn’t have to worry about hitting a wall with their data footprint growth. The data management should feel limitless.
- No virtual server maintenance – Users shouldn’t have to think about security patching, upgrading dependent modules, or monitoring virtual servers—all the tasks required to support 24 x 7 server uptime.
True freedom in the cloud is when you stop managing not just data center hardware, but virtual servers as well. Make serverless the true north in your cloud adoption journey.
Sign up for the free insideBIGDATA newsletter.
Speak Your Mind