BYOD and Cloud Apps don’t require a move to 90′s IT
In my last blog, I argued that Private Cloud vs. Public Cloud was the wrong debate, particularly in a Cloud App world.
Here’s why.
Some vendors will explain that the only way to preserve control over data is to go back to fully on-premises software. Not only is this untrue, but the private cloud “circle the wagons” approach regurgitates many of the things you hate about enterprise software:
- Difficult to implement
- Slow to change
- Inflexible
- Lack of innovation
As Ray Wang of Constellation Research put it: Enterprise software sucks. Rather than locking everything down and going back to 90′s IT, think about the ideal scenario:
- What if application rollout could be self-service for end users, IT provisioned, or a mix of both?
- What if the application user experience and features could improve at the speed of many of those consumer cloud services you use, while all the user data was still held safe and secure in your own four walls?
- What if you didn’t have to deal with scheduling and managing patches and updates, but allowed the application handle the bulk of it?
- What if the application scaled for peak times without you having to provision extra application resources?
Hard to believe? No doubt, it’s a different way of thinking, but it’s an approach Oxygen has had all along:
- Decouple the application and presentation from the data – this allowed us to innovate and add dozens of new features over the last few months with minimal IT overhead
- And, keep the user credentials and data under firm customer control and ownership – avoiding the lapses and security questions that have hit the cloud file sharing market repeatedly
Mobile file access doesn’t have to be a monolithic application hosted entirely in the cloud. Nor does it have to be a monolithic application hosted in your data center. Cloud applications can be built to provide continuous innovation AND control. You just need the right platform and approach.
-Leo
(@lleung)
Filed under: enterprise adoption, Enterprise storage, file sharing, Hybrid Cloud, Private Cloud, Private storage | 2 Comments
Tags: enterprise software, hybrid cloud
Authors
-
Leo
Let’s think for a moment how “Enterprise” IT ended up in a mess with such poorly written applications. We used to have mainframe, where everything was enrobed in the control of a single platform, with usually one database (DB2 or perhaps IMS) and one application layer – CICS. Then came the 1990s and with it distributed computing. Unfortunately the benefit of more pervasive enterprise computing also introduced less rigorous programming and management standards. For example, backup in the last 15 years has been the way many organisations have implemented archiving.
What is there to suggest that a change of platform will change the behaviour of these sloppy developers?I’d like to suggest the answer is none. Applications programming has a chance to be as badly managed “in the cloud” as it is in-house.
The real discussion is infrastructure management. IT organisations within large companies are poor at infrastructure management for many reasons, including cost/politics/business issues and probably more than anything, the historical position of having to deal with many years’ worth of legacy systems and hardware.
There are two issues to deal with here;
1. Start from scratch and build service-based IT offerings based on business need. Integrate these with cloud computing standards.
2. Make developers write proper code and create structured applications that work within a service-based infrastructure delivery network, rather than pieces of monolithic hardware.
We need both of these things to deliver the next generation of IT services in the Enterprise
Chris,
Thanks so much for the thoughtful comment! History and politics (sadly, the wrong influences) do indeed have a lot to do with the state we’re in.
While I completely agree that a two-sided approach is needed, I would argue that cloud app developers (at least some of them), have indeed built apps that work within service-based infrastructure delivery networks. While it’s often hard to tell the exact provider that’s being used, when you look at modern cloud apps, you can see they’ve built their apps as consumers of services, not physical or virtual “stacks” of database > SAN or FS > NAS. Instead, they consume the proper granularity of service to serve the application and business need. From fast simple databases to slower eventual consistency blob stores, you can build a cloud app that is completely service-based.
The problem is that they have often developed them to one infrastructure: Amazon’s. While we’re appreciative of AWS’ capabilities and innovation, this approach doesn’t really help many enterprises. The open question here at Oxygen is whether someone can build “service-based IT that integrates with cloud computing standards” without wrecking havoc on data centers and existing infrastructure. For example, while I personally worked on an object storage platform, I can certainly understand the resistance to having to adopt a brand new storage platform in my IT stack simply in order to leverage cloud apps.
Leo