Migration is already underway, but some of the world’s largest organizations are still reluctant. They handle proprietary data and a staggering volume of transactions.  They want to marshal information and deliver nuanced results.  But a lack of appreciation for the Cloud’s promise, together with questions surrounding security and cost, make many CTOs and COOs cling to legacy systems.

shutterstock_110126165_778x360

At the surface, the Cloud is seemingly straightforward even for occasional users; using encryption and/or password protection, individuals log on to a website, smartphone app, or similar “thin” interface. They may send an email, manage ‘Internet-of-things’ controls, reserve a seat, store a document, access a data base, or subscribe to an online tool for accounting or lead tracking.  While it’s clear that the “back office” intelligence sits someplace else, users probably don’t appreciate the architecture.

Understanding the Cloud’s capabilities and shortcomings is key to a thoughtful migration strategy. Can organizations really save money by replacing their on-premise solutions?  How do they comply with regulations around the world addressing the collection, use and transfer of personal data?  When migrating, do they retain control of their systems?  What about cyber security?  Here’s a simplified description for non-technologists!

Physical Server Farms and Virtual Capacity

The Cloud consists of millions of interconnected components and specially-configured resources that leverage them. By negotiating dark and lit fiber agreements, rights of use to subsea cables, data center leases, and related managed services, my colleagues and I help assemble the hard-wired network of computers and other facilities underpinning the Cloud.

Sophisticated programs pool capacity and various capabilities across dozens, hundreds or even thousands of servers to enable a virtual computing environment. The “-aaS” moniker – Software-, Platform- or Infrastructure-as-a-Service – denotes which combination of resources (among, for example, networks, servers, storage, runtime, applications) an organization manages for itself and which are outsourced.

The “Public Cloud” constitutes such outsourced capacity. An organization avails itself in part because the Cloud Provider does the worrying about how devices and connectivity within its sphere are employed.  An organization can self-provision such resources, facilitate broad access, and scale for volume-based offerings or other needs.  High-speed Internet access and ever-friendlier dashboards or other web page interfaces enable organizations and their users and customers to tap in.

But the Cloud’s great promise invites precautions too. For example, cheap and plentiful storage by big-name Cloud Providers should not lull focus from security and privacy measures and associated risks.  Similarly, while the Cloud promises that organizations can work with their data seamlessly and remotely without necessarily having to think about the placement, cost, or other characteristics of the hardware, country-specific data regulations override and can quickly draw back the focus to the physical location of the data.  And flexibility in employing capacity based on changing levels of need may be curtailed if an organization is subject to minimum contract terms or early termination charges.

How is the Cloud Used?

Ubiquitous storage and capacity is utilized in different ways depending on the program or task. Wide dissemination of information from a single source may be important.  A desired result could require constant updates to multiple databases.  Sometimes complex analytics are involved.

For instance, an organization must figure out how best to reach its users. Content is placed on one or many servers, depending on the popularity of the program and where it is accessed.  By using multiple servers in different locations, an organization can balance meeting the demand for content without having any physical presence in the locations where the users reside.  Remember the weather feed on nearly every website you visit?  The Cloud allows a weather website to accommodate billions of daily global hits by provisioning capacity and disseminating the latest local forecasts according to user browsing patterns.  Similarly, television or cable companies can efficiently stream live fútbol or cache content for video on demand close to users.

Collection of data is another hallmark of the Cloud, even if it’s also a challenge given the security and privacy concerns as identified below. Credit card billings, toll gate readings, stock trades, website visits, security incidents, package tracking and other commercial transactions generate enormous data.  Systems are increasingly being trained to “phone home” so that analytics can detect trends and predict outages.  Newer generation aircraft engines, for example, produce terabytes of operational details per flight!  The Cloud enables an organization to disperse its worldwide data processing and storage presence and thereby collect information for later processing.

The Cloud lets organizations harness additional tiers of capacity as needed to accommodate testing, staging, and other complex tasks requiring high throughput. For example:

  • Airlines – Carriers routinely run complex programs to manage revenue, booking, and flow of equipment, fuel, catering, crew, passengers and baggage. A blizzard or afternoon thunderstorm require even more capacity as algorithms reassign resources and notify vendors and passengers.
  • Automakers – Self-driving cars will run based on real-time processing of coordinates, landmarks, signage, road conditions, traffic and other inputs.
  • Broadcast – According to market research and consulting firm Devoncroft Partners, the Cloud reduces dependences on compression and per-channel facilities and streamlines transcoding, editing and playout.
  • Healthcare – Providers are increasingly reimbursed based on successful outcomes, benchmarking diagnoses, medical history, current medications, interventions, complications, and readmissions across entire populations.
  • Manufacturing – The Cloud has redefined product deviation analysis, brought new precision to just-in-time inventory, and effected infinite simulations for fluidics and wind.
  • Pharmaceuticals – Drug researchers now simultaneously run innumerable models with the human genome to discern the meaning of mutations, pathogen characteristics and cell susceptibility. The data can also hone precision medicine.
  • Retail – Using proximity and other management capabilities, stores know customers by analyzing their preferences.

In short, the Cloud permits an organization to exercise control of vast resources.

Migration Considerations

Organizations already imagining themselves in the Cloud navigate the important concerns associated with a migration strategy. Among numerous advantages, there are sensitivities and pitfalls too.

  • Cost – The Cloud carries an alluring narrative about cost savings. Naturally, enjoying true savings depends on numerous factors. On the one hand, old systems can be a drag on budgets. Organizations have sometimes provisioned the maximum required capacity even when it is rarely exploited. Equipment breaks or becomes outdated. Non-standard systems demand specialized personnel and custom tools. But it can be amortized and, like anything else once bought and paid for, it is bought and paid for. On the other hand, a Public Cloud theoretically allows a customer to ‘pay as you go,’ which is great insofar as it goes. However, the customer keeps paying. And paying for only the needed capacity may be belied by the sheer difficulty of managing it; savings may be mitigated by minimum commitments in Cloud contracts. It should be noted that while efficient use of the Cloud may reduce costs, it also reshapes a company’s financials — ongoing operating expense for usage fees, licensing and data center leases could be reflected in a noticeably new line item.
  • Security – For specialized data integrity or auditing needs, an enterprise may stick with a Private Cloud, which could consist of infrastructure deployment in leased raw data center space or an outsourced deal through which a Cloud Provider hosts and manages dedicated servers. But users of both Private Clouds and Public Clouds maintain cybersecurity by choosing authentication protocols, encryption levels and firewalls. Cloud Providers provide physical security, partition capacity to enable compliance with data laws, and facilitate monitoring of vulnerabilities. To be clear, they rarely guarantee data security. But even if not definitively, their offerings are increasingly viewed as more secure than individually-managed hardware. First, they can point to prominent reference accounts that require compliance with PCI credit card security standards, HIPAA security rules, and other common security regimes. Second, given that security is the biggest migration concern, many take comfort that competition among Cloud Providers drives such providers to make continual improvements.
  • Redundancy – A key risk that has evolved for organizations – especially for public companies – is whether they have adequate diversity and redundancy to ensure preservation of mission critical data. Data centers suffer outages. Cloud Providers and their customers should maintain failover solutions for disaster recovery by facilitating backup and parallel connectivity.
  • Scalability – The beauty of an outsourced solution is that resources can be quickly added or subtracted depending on the need at that moment. For example, organizations that maintain their own data on-premises may quickly need thousands of parallel computing cycles for crunching huge sets of variables. Again, minimum capacity requirements or other contract restrictions could hinder such nimbleness.
  • Service Levels – Cloud Providers offer guarantees within the range of most users’ fault tolerance and latency sensitivity. Larger organizations will find that SLAs are negotiable. In any case, organizations should keep in mind that remedies are often limited to credits against monthly recurring charges. Those with unique requirements may maintain certain functionality locally or set up different criteria in a hosted Private Cloud.
  • Physical Arrangements – While hosted Private Cloud arrangements draw on many of the same benefits available with Public Cloud offerings, data center leases demand their own attention. To avoid hidden costs, Private Clouds need to take a smart approach to determining fitting capacity, power charges, cooling, security, and connectivity. Leasing also involves allocation of expenses and risk, expansion options, and assurances of capacity portability and continuity.

Next Steps

Organizations don’t necessarily take the plunge – a thoughtful approach using available resources is important.

The evolution of Cloud computing is instructive for an organization’s planning. Historically, offsite capacity was simply used for storage, retrieval and backup.  Software-as-a-Service began as a way to alleviate the need of customers to download and maintain large software files in numerous locations.  When web transactions became rote and repetitious, the Cloud’s breadth became an obvious means to manage them.  Similarly, organizations can start their migration with the most common or resource-intensive functionality.

Incremental migration still takes time. Organizations must account for the possibility of long-term arrangements with providers and a variety of contractual contingencies that could limit their flexibility in making a quick change.  They also need a plan for defining security parameters, initially moving vast stores of data, and balancing workloads.

To avoid reinventing the wheel, there are plentiful programs to ensure proper back-office functionality. Vendors can help organizations run popular software, configure databases, move data between different Public Clouds, and leverage in-memory capabilities for better computing.  Off-the-shelf dashboards facilitate the control of systems virtually and economically.

No matter the industry, those seeking virtual availability of content, high-throughput computing, and global connectivity need to understand migration mechanics and issues.