Cloud on-boarding – Large Enterprises

We looked at some basics of the cloud on-boarding in part one of the series. This second post aims to un-cover a very high level view of some of the challenges which large scale enterprises face when on-boarding to a cloud based solution (mainly from the perspective of their collaboration & productivity tool sets, and less from the native operating system upgrade lens). This is essentially a mind set shift in how organizations will need to start thinking in terms of productivity solutions as a ‘service’ instead of a purchased and installed software which needs to be refreshed by their IT every X years.

In the 10 years from 2007 to 2017, the overall cloud market has gone from under $5B to upwards of $130B, this growing trend is only heading in one direction ! This when combined by the fact that hybrid & private Cloud are starting to shrink with a negative growth of 4 & 5% respectively (State of the Cloud, 2017 report); leads to a logical conclusion – IaaS & SaaS are going to be expanding in the next foreseeable future. Lack of resources/expertise, the #1 cloud challenge in 2016, was less of a challenge in 2017 with only 25 percent citing it as a major concern, down from 32 percent in 2016. Concerns about security also fell 4% to 25% (of the 1002 companies surveyed*). These trends indicate a sharp rise of interest in cloud experimentation, and alignment towards strategic cloud adoption.t text

With this context, i would now like to focus on one other aspect before we get to the meat- An uber strategy that enterprises can use to help accelerate their drive from watchers to adopters. At a very high level, the below steps needs to be accomplished – This is definitely not comprehensive & in depth, if fact, it’s cursory enough that each one of them can be an entire book in it’s own;

• Understand and get convinced by the Cloud value proposition that goes beyond the cost savings, e.g. business focus, security, scale, availability & increased resource awareness.
• Recognize one of more executive sponsors & generate organizational buy in.
• Initiate platform consolidation, including but not limited to asset management (inventory, categorization, future modeling.) & Line of business apps.
• Start exploring the cloud providers, selecting one is going to be one of the biggest gating factors, especially if a large scale enterprise chooses anyone not in top few providers.
• Internal re-orgs, development process changes, re-tooling of company apps & deployment tools, IT restructuring & such. Not all of this needs to happen or happen as a pre-requisite but attention should be paid to identify blockers e.g. if the lead generation app relies on hard coded versioning of MS Word / uses legacy smtp stack for emailing; they might need change. Hire experts if needed.
• Identify the support and monitoring structure.
• Set user expectations, update support guidance documentation.
• Understand how the move might impact the end users – what are the key scenarios your users rely on ? How will this move impact them? Execute the on-boarding.
• This is where the cut over happens.

If the process seems daunting, then i must caution that a fragmented / broken experience for users defeats the purpose; so my goal is not to scare but to ensure that the process is well understood and done with back up plans & appropriate risk mitigations.

Now the meat, in my 3 years of dealing with multiple large scale enterprises, here are the top few of most common issues we saw.

• Existing network configurations – Network piece of things is the toughest to execute on(Per Gartner on Office 365), not because of the provider though but because of the existing firewall rules, traffic limitations, static network configurations already in place at the customer endpoint. Customer first mile is not used to the kind of traffic a mature SAAS app like Office 365 can generate, it is not built (read: lack of resiliency, latency & throughput) to handle it and must be extensively discussed and worked upon, before embarking on cloud adoption journey.
• Identity management & providers – This is the key, users need to continue to authenticate as they have been, identity ( both for existing users and ones joining the vCloud directly need to be managed effortlessly.) This is one of the top support call generators. Choose your identity provider wisely. Preferably one recommended by your cloud provider, E.g. for Microsoft 365 cloud, we recommend Azure Active Directory due to it’s native support.
• Security – Security is a top concern ! One of the most common ways to address this is via, ‘Role Based Control’. Role based control to manage secure access to various layers of company data. This is similar is concept to the RBAC model of Microsoft Exchange. A conceptually similar model can be adopted by the company.
• Data protection & privacy – Ensuring to maintain data protection & privacy (for data both @ rest & in motion) using secure auth, encryption.
• Network Security – Establishing secure network peering, we at Microsoft 365, recommend Microsoft Azure ExpressRoute ( this is a great primer on it.) but other providers have equivalent network guidance in place.
• Over pivoting on automation – This one is pretty interesting, SRE’s and PE’s especially love automation. We do this at times, while knowing that this entire exercise is a one time exercise. Do resist your love for Investing too much in automation (E.g. Scripts) to enable on-boarding. Key here is to consider, at what point does my ROI start dropping / flattening.
• Business Continuity via Disaster Readiness – You need to keep the lights on. Failing to plan is planning to fail, is most apt in this case. Failing to plan for BCDR (Business continuity & disaster) and also, failing to ask your chosen provider to provide key data & tests on BCDR processes. All providers including AWS, GC & Azure are prone to failures.

Top Few Caveats to watch out for !

During the actual execution, it’s critical for the provider & the customer to stay on the same page. Weekly meetings are a great way to enable that, so are monthly leadership calls. The journey is a partnership & not a tax. In an ideal case results are not going to be magical, but previously understood, well tested and fully deterministic. Ambiguity at either end will be a time sink – It’s imperative that the provider has a custom created, actionable and well documented process documentation. It needs to be tailored to the customer, e.g. that has their unique endpoints, custom configurations, IP Addresses & such. A detailed project plan is another asset to have.

Setting right expectations at all levels in the company – This is another blunder which i have seen more than one company make. Managers assume a certain outcome / path without reality check with the provider or vice versa, where the provider assumes a certain level of support. Both of them resulting in precious cycles wasted.

Combining other streams of work. This was one of the issues we kept facing with a key customer of ours, wherein they decided to upgrade their user machines, operating systems, mobile devices, MDM solution, ALL while trying to on-board to cloud. This kind of approach, can sound clean from one perspective, but can cause issues like-

◦ App incompatibilities – Internal apps, especially legacy systems are hard to fudge to integrate with your shiny new cloud system. Example: For a big pharma’s migration, we realized 2 weeks before cutoff, that their HR system which generates pay stubs will stop working because of missing FAX integration !
◦ Circular dependencies – your MDM solution depends on a client version which is not compatible with the version of the core cloud infra service.
◦ Support and help desk overload.
◦ Timing issues – the version of ‘one thing’ is near end of support but cannot be updated because of circular dependency on another.

Footnotes-

* – The split among those 1000, was near even 517 Large enterprises & 485 SMB’s.

Thoughts ? Suggestions ? Feedback ? – please let us know via your comments.