Digital transformation. Software-defined business. All companies are software companies. Webscale IT. We have lots of ways of talking about the reality today that companies must have IT people and systems that can reinvent the business. Companies that don’t will fail, even if it’s slowly.
Webscale IT and webscale infrastructure have an especially strong focus – across organizations of all sizes. Webscale principles came out of the biggest of the online companies, including Facebook and Google, as they applied cloud-like technology and processes to their internal systems. Webscale delivers far more than “big” – most importantly, it delivers resiliency. Pieces can fail and the system as a whole will not.
A lot of technology already benefits from webscale architecture. OSes, networks, the web tier, and application servers all leverage webscale properties, with elements like multi-threading, redundancy, and web load balancers. Even databases have achieved webscale properties, with replication and secondary servers providing that much needed resiliency. The problem is those webscale properties have failed to benefit the applications those databases serve.
Why? If the database runs at webscale, doesn’t that mean the apps benefit? Typically no, because the app has to be directly connected to the database servers, and so even though the database has great failover properties, those disruptions impact the app.
The database, no matter how sophisticated, cannot deliver webscale on its own. In fact, the more sophisticated the database architecture, the more database coding has to be in the app. This approach is not sustainable – organizations cannot constantly custom code their applications to take advantage of database capabilities.
To achieve webscale at the database tier, organizations need to break the 1:1 tie of apps to databases – they need an abstraction layer. That architecture – with software sitting between apps and databases – enables the resiliency that’s fundament to webscale principles. Only with that abstraction layer can the application be shielded from disruptions at the database tier.
The abstraction layer approach serves a number of use cases:
- failover that’s transparent to the app, within or across data centers
- zero downtime maintenance
- cloud migration
- database upgrades, especially scale out
In the next series of blogs, we’ll explore those use cases in more detail and highlight case studies of customers benefitting from them. You can get a sneak preview by watching our CEO Justin Barney talk through these use cases.