Data creation and storage are growing exponentially, and more organizations are turning to managed service providers (MSPs) to handle the infrastructure needed to process it.
A survey of nearly 200 IT leaders conducted by DataStrike, a provider of managed IT services, finds while only 19 percent work for organizations that outsource the management of data infrastructure, more than three-quarters (76 percent) would consider relying more on an MSP for this function.
However, among the 19 percent that already work with an MSP, there doesn’t appear to be much customer loyalty. Nearly three-quarters (73 percent) said they would be open to working with an alternative MSP.
Opportunities and challenges lie ahead
Clearly, there is much work to be done to increase MSP’s share of the overall total addressable market for managed data infrastructure services. The survey, for example, finds that even though more than half (54 percent) of respondents admit they lack resources and tools to meet the growing demands, well over half (55 percent) continue to rely solely on internal IT teams to manage their data infrastructure.
Specific areas where IT leaders are looking for additional expertise include optimizing data infrastructure costs (23 percent), optimizing cloud computing costs (23 percent), and developing a data infrastructure strategy (21 percent).
Scalability vs. infrastructure limitations
It’s safe to assume that advanced analytics coupled with artificial intelligence (AI) will be a major factor in the data infrastructure strategy most organizations are still trying to define. Many are quickly realizing that the legacy IT infrastructure they have in place is not going to be able to scale to the level these types of applications will require.
Additionally, most organizations are still working to determine the best environments for running AI inference engines. It makes economic sense to rely on cloud services because of all the data and graphical processor units (GPUs) needed to train AI models. However, the inference engines an AI model creates for performance and compliance purposes should be deployed as close as possible to the data being processed and analyzed in near real-time. Much of the data already resides in an on-premises data center or at the network edge.
It’s unclear which types of processors suit running inference engines best. The ongoing GPU shortage and their high cost make it harder to rely on them. Other alternatives might offer more cost-effective options.
Why MSPs will be key to managing AI and data infrastructure
In addition to acquiring the expertise required to successfully manage those advanced analytics and AI workloads, it’s clear MSPs will need to spend some time and money convincing organizations that they are a viable option. There is a lot of historical bias toward managing data infrastructure internally. At the same time, however, many of those same internal IT teams are under pressure to enable their organizations to achieve their AI goals as quickly as possible.
Ultimately, it’s only a question of time before most organizations begin to rely more on MSPs to manage data infrastructure. The only issue to resolve is determining which MSPs can provide those capabilities at a cost that is less than what it might otherwise cost those organizations to manage that infrastructure themselves.
Photo: StudioRomantic / Shutterstock
