Migrating To The Cloud
There’s an understood expectation that enterprises realize immediate economic benefits when they migrate to the cloud. The unstated value, however, is a renewed focus on their core business and applications. For most users, this translates to dynamic capacity planning as well as improved time to market for new applications deployments.
But enterprises should expect more. Enterprises are now only scratching the surface when it comes to understanding the “black box” that is the cloud. The true value of the cloud is only realized when it allows your enterprise the visibility and the control at every layer of the IT stack.
From server closets to in-house data centers. By the late 1980s, enterprise IT infrastructure was in a closet at the back of the office. But as computing capacity increased, so did demand, and soon employees were asking for services like email and access to company databases and shared files from their desktops. At the same time, the IT infrastructure itself was maturing, enabling a more advanced type of enterprise-wide IT. So IT moved its infrastructure to a dedicated in-house data center.
Data center colocation. Fast-forward five years and the infrastructure that was state-of-the-art when the enterprise bought it is now outdated. New IT infrastructure is more capable, but also more complex — and more expensive. So enterprises began to “outsource” their data centers needs. Enterprises began the process of turning over management of the data center to colocation providers but continued to own and control their own hardware; they still bore the costs of server management and administration (OPEX), which continued to increase — both in absolute terms and as a percentage of total data center costs.
To reduce CAPEX while also addressing increasing OPEX, enterprises need to rely on an intelligent control platform that integrates the IT stack from the physical to the virtual. It is only with increased transparency that enterprises can realize CAPEX and OPEX savings.
Cloud is the next step in this evolution. Enterprises are moving to the cloud rightly expecting:
- To realize CAPEX and OPEX savings. Cloud goes a step beyond the cost savings benefits of colocation with variable usage-based pricing. OPEX is further addressed in cloud with orchestration stacks and services for compute-storage-network that allow the management of many physical entities from a single logic point. Furthermore, when the cloud is in a massive scale data center, the provider should be realizing (and passing on) power and cooling savings from economies of scale. And because the cloud can “be” anywhere, there are geo-location savings opportunities as well.
- To focus on the business. Colocation gets the enterprise out of the business of building and running data centers, but not out of the business of procuring, installing, and maintaining IT hardware. Cloud does. Cloud allows the enterprise to focus on its core competencies and leave IT management to a provider who can — because cloud is their core competency — do it better.
- To quickly “right size” capacity. When adding capacity involves a request to the IT department to procure and install new vanity hardware, scaling quickly is very difficult to do. When the cloud provider already has scale built out, its customers can access IT capacity in only as much time as they can log into the cloud provider’s system and make the request. Often, the cloud relies on commodity hardware, which is both quicker and less expensive to procure. All that means developers can roll out applications themselves, bringing those applications to the business faster so the business can realize the benefits sooner. Furthermore, cloud enables businesses to add capacity for a period of time in situations where they couldn’t economically justify purchasing it outright (quarterly analysis on business data, for example). That levels the playing field for small companies to compete with their moneyed competitors.
CAPEX and OPEX savings, a focus on the core business, and on-demand capacity additions are indeed benefits that enterprises should expect from the cloud. But they’re not the only benefits enterprises should expect. The true value of the cloud is only realized when it provides more transparency, more visibility, more control, and more insights than the enterprise had with a colocation provider or on-premises data center. With that transparency, visibility, control, and insights the enterprise should be better able to make better decisions.
Yet in many cases, migration to the cloud does the exact opposite. In many cases, the cloud becomes a black box. Enterprises send their applications and data into the cloud knowing little if anything about how it is secured, powered, and managed. “Since businesses no longer manage their servers directly and the people who manage the collaboration platforms may be located somewhere else in the world, the need for visibility into how data and intellectual property are being managed has become even more business critical.1
“The complexity of the environments will increase with increased virtualization and adoption of the cloud,” explains Deepak Satya, head of data center services for Wipro Technologies.2 So without visibility into metrics at every layer of the IT stack, the enterprise won’t know the impact of changes in one layer on the others. For example:
- Most enterprises have business process applications (e.g., ERP, CRM) running on a database. The database can run on a storage infrastructure that is either disk drive or solid state-based. Having visibility into the key storage metrics (defined below), enables the enterprise to optimize policy decisions about where the data should best live.
- At the server layer, visibility into key metrics allows the orchestration layer to choose the optimum execution venue for any given workload given target cost and runtime. The same workload depending on time of day/energy cost/user demand can then be live migrated to the best execution venue.
- At the data center layer, visibility into power and cooling metrics enable the orchestration layer to adjust workloads to optimize for a given cost of running the application target. Visibility into infrastructure condition metrics (is any part running in degraded condition?) enable applications higher up in the stack to live migrate or spin up additional capacity.
Understanding those intertwined impacts requires visibility at every layer of the IT stack.
At each layer of the IT stack, there are key metrics that the enterprise should expect to see from its cloud provider:
- Disaster recovery (DR) readiness
- Run rate cost
- Response time to queries
- Who is talking to whom?
- Applications being run
- Enforcement of isolation policies
- Throughput/latency hotspots
- Hotspots for wear leveling and predicting failure
- Loading in terms of CPU cores and memory in use
- Efficiency of workload execution
- Current, trend, and peaks in power and cooling
- Infrastructure condition
Beyond the transparency just described, the enterprise should expect insights that enable optimization of its IT. A cloud provider should be sharing intelligence and knowledge with its users — based on the provider’s deep analytics and simulations — to help them make informed decisions. For example, recommendations about capacity planning (at the application or infrastructure level); where to best place workloads (including impact to other systems); and sustainability recommendations based on power costs and availability. Insights that allow users to improve their applications shouldn’t be reserved for only enterprises with deep pockets.
By moving their infrastructure to the cloud, enterprises can realize benefits that extend beyond the economic value coming from IT out-sourcing. Those benefits depend on a cloud that is transparent, on a provider that offers visibility and control over applications and on in-telligence, which comes from real operating data and metrics. Enterprises should expect their cloud provider to offer insights that will help it squeeze every ounce of value out of the cloud. That, after all, is the point.
Enterprises should expect more from the cloud. And get it.
- Wired Innovation Insights, “A Cloud Platform Without Visibility is Like Driving Blind,” June 6, 2013.
- WIPRO, “The Road to Better Visibility: Virtualization, the Cloud and Managed Services,” September, 2010.
This article was originally posted “Migrating To The Cloud” from Cloud Strategy Magazine.