If you’re an enterprise, and you’re an IT guy or gal, here’s a business view of your data center :
- One part is stagnating, and will slow down (existing lines). Pressures are piling to keep this part of the business lean and mean. But the revenue it generates is still humongous and dwarfs the digital line tremendously. The IT stack here consists of the familiar stacks of the last decade – 3-tier apps, highly reliable systems, SQL databases, if-it-aint’-broke-don’t-touch-it philosophies.
- Another part of your business is growing and needs to grow (digital line). This is spawning a flurry of IT activities with new IT stacks – digital apps, new middleware, new infrastructure services, new IT SLAs, new demands. However, the revenue contribution still pales in comparison with existing lines of business.
In enterprise architecture thinking, the most important principal (IMHO) is continually aligning IT imperatives with business trajectory. Otherwise, IT will quickly lose relevance.
Let’s focus on the first part – the Existing Lines. Here’s Mike’s Take on ways to keep your DC lean and mean, with a modern data center architecture.
- Reduce cost by reducing the absolute infrastructure footprint through all-flash data centers. Today, we have moved past the intersection point where flash is more cost effective than disks. Flash brings about multiple-factor compression capabilities.
- Offload cost by moving cold, stale data to public cloud services. Do this through cloud-integrated / cloud-enabled software like Cloud Tier, Cloud Pools, Cloud Array.
- Use these seamlessly with multiple public cloud providers so that you create procurement choice and cost leverage.
- Avoid the cost of infrastructure ceilings by adopting systems with scale-out architecture and design. Moving forward, start at the right size and grow seamlessly as needed. Apply this to flash systems as well.
- Avoid the cost of down time and slow time by building a trusted DR and replication that can adapt to changing production traffic. Swap between active-active, sync and async.
- Avoid the cost of auxiliary infrastructure sprawl. Auxiliary systems like backup servers, software and media servers can be eliminated with software-defined technologies like direct storage-to-backup. Check out ProtectPoint.
- Finally, eliminate DC facilities costs when you drive a modern data center with optimization capabilities described above.
It should look something like this :
(ps : the images don’t represent the actual number of physical boxes 🙂 it’s drawn this way to surface the areas of optimization that is possible. Cheers)
So, where’s converged infrastructure in all this DC optimization? It is definitely a core component in a modern DC. In the next two articles, I’ll address how you can put these blocks together to make migrating to the new architecture effectively.