Should you outsource your data centre?

Kathie Lyons EVP & GM of ParkView and Entuity, Park Place Technologies, discusses how organisations should decide whether or not to outsource their data centres, and when SD-WAN becomes a good option

Some organisations like to outsource technology challenges like data centre operations, preferring to focus on their core business. Others like to have direct control over their infrastructure. To outsource or in-house is an age-old technology conundrum, and with new cloud capabilities being introduced seemingly every day, it's a question which isn't getting any easier to answer.

Computing caught up with Kathie Lyons, EVP & GM of ParkView and Entuity, Park Place Technologies to discover her views on how organisations should go about tackling the challenge.

Computing: How do organisations choose between operating their own data centres and outsourcing them?

Kathie Lyons: The organisations that succeed with outsourcing, first assess their overall priorities and strategy for their IT ecosystem, then map those to available skills, resource capacity, and budget. Going through this assessment correctly, will highlight viable outsourcing opportunities. Management can then clearly articulate the achievable and measurable objectives that can be met through outsourcing.

Measurable objectives would include cost, resource optimisation, streamlining operations, and competitive differentiators. The assessment facilitates the decision-making process to keep a function in-house or outsource it to a trusted partner. Capital cost avoidance and operational cost savings are frequently the primary drivers so that new strategic programs can be funded. Savings can be achieved by right-sizing the cost of skilled resources, services, and technology via outsourced managed services. This frees up highly skilled and compensated internal capacity to focus on the transformational activities they are best suited to undertake.

Hardware refreshes should be limited to critical hardware to assist in changes to the working environment, such as working remotely or implementing new products/services. Third-party maintenance options can extend the lifecycle of storage, server and network hardware post OEM warranty/support. You can also consider purchasing pre-owned hardware, when additional capacity is required or even gain additional revenue from a secure IT Asset Disposition (ITAD) programme, if necessary, too.

Cost-control adjustments are important; first, to ensure the key vendors' health; and secondly, to utilise partnerships to create win-win opportunities. For example, perhaps a managed service provider or third-party maintenance provider will allow greater flexibility in service level agreements. You may even be able to alter the SLAs by location when an OEM may not be able to do so.

Partnering can often provide new cost-saving opportunities when both parties tackle a challenge or goal with a win-win mentality. Outsourcing can reduce the time-to-value of advanced technologies by avoiding significant internal training or hiring while enabling current staff career enrichment by using automation and outsourcing to eliminate commodity tasks (e.g., discovery, patching). Executing these low-level foundational tasks through outsourcing allows an IT team to execute the strategic decisions that the data reveals.

CTG: What are some of the benefits of running them yourself?

KL: Certainly, there are advantages to running your own data centre, especially if you have defined your company's differentiating characteristics and aligned with those characteristics across the IT ecosystem. Owning all or some activities in the data centre enables control and governance to be managed within your organisation, which may equate to greater flexibility and centralised decision making. By keeping the work in-house, you can retain flexibility in work assignments and schedules. However, it's a double-edged sword because flexibility can lead to missed delivery dates and unnecessary reprioritisation exercises. IP protection and data security are cited as reasons to keep a scope internal. Strong security policies and processes with a trusted partner must be reviewed and tested to verify execution.

Many businesses choose to manage their own data centre operations because they believe they have unique processes that would be difficult to transition to a new team. An outsourcing assessment forces the re-evaluation of internal processes/practices. It also identifies opportunities to standardise on common best-practices. This is a worthwhile undertaking, whether data centre management is internal or outsourced. There may also be a cultural mismatch with an outsourcing strategy if a company's guiding principles include keeping innovation and technical expertise locally to retain and protect institutional knowledge. In this case, outsourcing onshore vs. offshore would be a better fit (which would offer larger savings).

Tapping outsourcing partners can access expertise on an as-needed basis, without paying full- or even part-time salaries. It's worth asking what data centre monitoring, maintenance, and optimisation tasks can be handed off to outside experts or even computers. With the rise of machine learning-based systems, for instance, you may not need the same number of administrators sifting the daily influx of network alerts as you did just a year ago. Another major advantage of a third-party maintenance plan is that an EOSL date means nothing to the provider. Third-party maintenance providers can offer support for systems well beyond the OEM EOSL date, meaning that an organisation can use a performant system for its full life. An additional bonus is the reduction of migration interruptions and effort.

In some cases, a company will choose to use hardware through the initial warranty and with an extended OEM warranty, only to end up replacing a fully functional, and possibly business-critical machine, because of an arbitrary EOSL announcement. It is important to realise that the EOSL date does not, in any way, indicate that a piece of hardware is no longer relevant. Instead, it is a sign that the OEM no longer benefits from offering support and wants to focus on newer hardware. While switching to a third-party maintenance and operating system support plan is possible past the EOSL date, there are often benefits, besides cost, for using the service in the place of an extended warranty. Legacy hardware that still plays a key role in enterprise operations can be a double-edged sword. On one side, it can meet performance demands while providing incredible fiscal savings compared to a premature refresh. On the other, it can be more difficult to manage, upgrade, and maintain because it depends on legacy parts. Working with a third-party provider deals with this second edge of the sword through a combination of data migration, hardware transportation, and operating system support services. As a result, it is often best to turn to third-party maintenance services well before the EOSL date.

CTG: Why is data centre maintenance important, what sort of things can go wrong?

KL: Traditional IT infrastructure maintenance or "keeping the lights on" doesn't get the attention compared to business transformational initiatives. It definitively should because it's the lifeblood of most companies. You don't want to learn the real value through an infrastructure failure that threatens your opportunities - or your existence as a business. IT infrastructure is the foundation that all technology services are built upon. A significant failure of HVAC, power, storage, servers or networks could result in prolonged downtime for the businesses and customers you support. ‘Regular hygiene' items such as OS patching may seem mundane and unimpactful, but if allowed to lapse, security breaches, regulatory/legal compliance violations, or downtime events may ensue.

Gartner pegs the financial costs of downtime at $5,600 per minute, or over $300,000 per hour. And a survey by IHC found that enterprises experience an average of five downtime events each year, resulting in losses of $1 million for a midsize company to $60 million or more for a large corporation.

The time spent recovering can leave businesses with an "innovation gap," an inability to redirect resources from maintenance tasks to strategic projects.

Sustained performance and continuous availability are taken for granted because IT leaders are focused on delivering their performance and SLA goals for mission-critical applications. Corrective actions in response to degraded performance or lack of availability take precedence over all other work. You must choose maintenance providers with the skills, resources, parts, and tools to provide immediate resolution to minimise any impact. Acceptable SLAs and compliance for full resolution are table stakes in the data centre maintenance game.

CTG: What's next for data centres, is SD-WAN set to take over?

KL: It is happening now. We have seen rapid adoption of SD-WAN across our customer base; COVID-19 certainly accelerated that trend. Customers had to monitor networks that had suddenly become highly decentralised, including many at-home edge devices. The jump to remote work will be a permanent one for most companies. IT Leaders have adopted an SD-WAN strategy to ensure monitoring and provide network services that more easily adapt to ever-changing workforce needs.

From the data centre to the CIO's office, nearly everyone is looking to increase flexibility and remote management capabilities as part of an upgraded resilience strategy. Having faced a sudden onset pandemic that challenged operations worldwide, the IT industry recognises that preparations can never be too robust. Throughout the crisis, SD-WAN stood out as one of the infrastructure choices that helped many companies survive and adapt.

As organisations rushed work-from-home arrangements into overdrive, that new model was often facilitated by SD-WAN. It's easy to see why. SD-WAN has long been considered a promising solution for companies with numerous locations. And it became an asset for more enterprises and SMBs as countless employee living rooms became new remote locations. From optimising utilisation of limited bandwidth to enabling a skyrocketing number of VPN connections, SD-WAN saved the day in many cases. Having witnessed this success, countless businesses that had delayed SD-WAN deployments are considering them now. Gartner predicted that 50% of routers would be replaced with SD-WAN by this year, and recent events could drive that number upward. The question with SD-WAN at this point is less "if" than "how." To mitigate the risk inherent in any technology implementation, some organisations are looking to stage their deployments, transitioning offices or groups on a prioritised schedule.

Another important consideration not receiving due attention in some cases is network monitoring and maintenance. Unfortunately, there is a perception in certain circles that SD-WAN provides the only monitoring dashboard the network operations staff will need, but nothing could be further from the truth. For one thing, multi-vendor SD-WAN is not unusual, and vendor-specific toolsets cannot provide full visibility in these circumstances. Moreover, network oversight must extend to LAN and WiFi and peer all the way into business application-related performance. More sophisticated remote management capabilities have also become essential when staff could be cut off from network assets due to public health-related restrictions or become unavailable in the case of illness or disaster. NOC personnel need to discover assets and visualise the network even as rapid changes are enacted, and automate key tasks to ease the management burden on limited staff. Troubleshooting SD-WAN policies and optimizing performance from a distance are among the other obvious benefits. It might be possible to make networks function with eyes bopping between dashboards. Still, it will never be efficient and can't be as effective as having comprehensive transparency, management, and analytics available in a single pane of glass. Should another crisis strike, robust monitoring and maintenance capabilities will likely prove themselves as important as SD-WAN itself did in the last one.

This article was sponsored by Park Place Technologies