The legacy cloud epidemic

The problem with “rented server” cloud upgrades

Upgrading “rented server” cloud core systems is a complicated process that can take years and cost millions in service fees. That’s because these core systems have many customizations that need rewriting for the new version. Also, when upgrades aren’t backwards-compatible, the upgrade process can require complex synchronization with other systems.

With such slow and expensive upgrades, most insurers eventually freeze their system version, forgoing new features and fixes. As a result, they continue using outdated and unsupported software which limits agility, slows product development and raises security risks.
To make matters worse, many deployments are already one or more versions behind before they’re even live in production.

How does a mature cloud platform make upgrades easier?

A mature cloud platform champions the philosophy of small, weekly updates without impacting client deployments or their uptime. This is achieved by running one version of the system for all clients with a public library of backwards-compatible APIs.

Vendors supporting multiple customized versions of their software must maintain multiple codebases which divides and dilutes R&D and support resources. Even many cloud-only vendors create customized instances for individual clients, resulting in similar issues.

A mature cloud platform takes a different approach with all clients running on the same version and codebase. It takes feature requests from all customers and makes customizations for none. This modern approach delivers industry-leading platform performance and accelerates R&D.
Frequent, incremental and backwards-compatible updates are the standard approach for high-reliability platforms like Gmail, Zoom and literally every major website. That’s why they keep working.

Resilient design for business continuity

• Data Center Redundancy: Built-in redundancy and multi-site failover allow insurers to operate without downtime even during localized outages. In contrast, “rented server” cloud platforms are highly fragile and subject insurers to dangerous business continuity risk.
• Microservice Redundancy: A mature cloud platform is composed of many stateless services running redundantly across data centers and automatically scaling to meet demand.
• Backups and Recovery: Active/Active [HB1] recovery type across data centers supports automatic failover, load balancing and robust disaster recovery capabilities. A backup strategy should also include monthly automated restoration tests; as seasoned CIOs like to say: “If your disaster recovery strategy isn’t tested regularly then you don’t have a disaster recovery strategy.”

Scale up or down to meet real-time demand

A mature cloud platform supports autoscaling which means automatically spinning up and shutting down servers in response to real-time demand. This capability allows core platforms to meet peak demand without sacrificing speed or reliability and later reduce capacity so resources aren’t wasted—all done automatically without system interruption.

Autoscaling is especially valuable during unexpected traffic surges such as:

• Open enrollment for group benefits.
• After catastrophes that result in a surge of claims.
• High activity times of day.

You can’t cut corners

Mature cloud infrastructure may be well understood by the likes of [HB2] Google, LinkedIn and Dropbox; but it still requires experienced technical talent and a significant investment. Conversely, installing legacy software on rented virtual servers is easy. When vendors cut corners, insurers suffer, and they may not even realize until it’s too late.

Here are some basic questions to answer when evaluating new enterprise technology:

• How often do upgrades occur?
• Is there system interruption during upgrades?
• Can I get an evaluation license?
• Does it scale automatically when demand increases?
• Are uptime numbers published?

Someday, high-quality cloud will be standard. But today it’s a mix in the insurance industry. Insurers must ask the right questions or their next platform might already be outdated. To make matters worse they’ll need another expensive system change sooner than they realize. Smart insurers make wise decisions and build for the future.

Dan Woods is the founder and CEO of Socotra, the first modern, mature cloud platform for technology-driven insurers. Before Socotra, Dan did IT [HB4] investing and portfolio management for Formation 8 Partners (now 8VC) where he specialized in data platforms and smart enterprise software. Prior to that, Dan’s depth of enterprise experience was honed over his six years with Palantir. There, Dan was an early member of the core engineering team and later transitioned to run Palantir’s partnerships as well as several deployments. He holds a MS in Computer Science from Stanford.

See also:

• The modernization of insurance finance starts with infrastructure
• Lemonade Insurance technical glitch exposes 190,000 driver’s license numbers
• Are your third-party claims providers cyber-secure?