When selecting Cloud partners, companies often obsess about vital statistics; how many, how big, how fast?
When you’re procuring for on-prem do you take this approach? Often not. You’re usually working hard to niche user/business requirements into budgets whilst simultaneously working on down-channel requests from your bosses asking about TCO, ROI and business case justification.
For reasons I cannot yet explain, once the Cloud is an option, many organisations lose sight of the imperatives and drivers behind the implementation of IT systems and fixate on the flashing lights, speed/recovery/SLA claims and the cool biometric technologies employed by firms providing datacentre and Cloud services.
Don’t get me wrong, I’m all for that. I like a nice datacentre tour as much as the next guy (or gal) but in reality, when looking to move into the Cloud there are some serious questions you need to ask your potential providers and partners that are beyond the standard operational elements that make up the headlines.
There are dozens of questions you might want to ask, but I’m trying to give a little focus here and based on my own experiences (as both a consumer and provider of Cloud and Cloud hosting services) I’ve boiled down to what I’m calling “The Big 7”:
- How can I get my data back if my plans change?
- Tell me about your <insert technology here> knowledge and experience.
- Does your contract include real-world skin in the game?
- Can you tell me at any time where my data is in a geo-replication scenario?
- Can I speak with existing clients of a similar size and complexity as me, aside from those you publicly case study?
- Can you run through how I will manage and manipulate network infrastructure?
- Can you walk me through your procedures for major incident handling, what happens in case of flood, fire or plagues of locusts?
Let’s look at these in a little more detail:
How can I get my data back if my plans change?
Moving into the Cloud is hard, moving out of the Cloud if you need to can be painful in the extreme. If you need to move workloads (back to on-prem or over to another partner, for instance), how would your partner facilitate this? If you need a big chunk of data in a hurry (perhaps to support a legal challenge) how would your partner make this happen? Ask your partner these questions as they will frequently have great answers for on-boarding but you’ll want to understand off-boarding before you take the plunge.
Tell me about your <insert technology here> knowledge and experience.
If you’re looking to have a workload managed (for instance SharePoint or Exchange) you will want to ensure that the partner has plenty of experience in this space. You’ll already know from your own experience that managing complex application workloads takes capability, experience and knowledge in varying degrees – does your partner fit the bill? If not, do they have a go-to partner (that you would work with) or can you engage your own choice of partner?
Does your contract include real-world skin in the game?
I’ve had many conversations with Enterprise level CTO/CIOs and they nearly all state that one of the reasons they hesitate in moving workloads to the Cloud (especially those considered mission/business critial) is the lack of “skin in the game” of many hosting/Cloud partners. In the context of hosting/Cloud this boils down to the risk share represented by financial penalties attached to failure to meet SLAs. If you run a website that drives $50,000 per day of revenue on a platform hosted in the Cloud and the service (for whatever reason) goes down for 2 days (unlikely I know, but bear with me) this has a directly attributable cost of $100k to your business. If you pay £30k per month for your services and the financial penalty is (for instance) “monthly cost divided by outage multiplied by two) then your compensation might only be $4k – so you’re $96k in the poop. If a Cloud hosting partner really believes in their service, capabilities and infrastructure they may well offer a much better claw back in their SLA.
Can you tell me at any time where my data is in a geo-replication scenario?
I’ve included this as it’s highly relevant to many business types – you may be able to put data in the Cloud in certain scenarios, but there may be a proviso that you know where it is at any given time. If your partner has datacentres in London, New York and Singapore – you might not be able to hold your data outside of EU and US territories – a problem if your partner’s geo-replication arbitrarily replicates to all three sites. Geo-replication is a double edged sword so make sure you understand how your partner employs and manages it.
Can I speak with existing clients of a similar size and complexity as me, aside from those you publicly case study?
This question is not really only attributable to the Cloud – you should always ask it, but ask it a litttle differently. “Friendly” clients are often used as references and they’re pump-primed to say the things you want to hear. Ask to go off beam a little and speak to clients that are not cited on web-sites or in literature or other materials. You’ll get a less polished response that is likely to be closer to reality. We all like the idea our new car does 45mpg, the manufacturer and an “independant” tester tells us it will do it – but have you ever actually managed to acheive published mpg figures?
Can you run through how I will manage and manipulate network infrastructure?
Outside of SaaS, the Cloud is much more than just applications or servers. You’ll likely need VPNs, firewalls, out-of-band management and connectivity to storage amongst other infrastructure elements. Ask about these. I recently worked with a client who fell fowl of a statutory audit as their hosting provider managed the firewalls in place and as the provider knew (and had access to) passwords for the firewalls, the customer was in breach of various security protocol requirements. Not a fatal oversight, but frustrating for them for a period of time. Many Cloud partners will give you full control of (for instance) virtual machines but very limited control of the plumbing around the VMs. This may not be something you are concerned about in the first instance but is something you want to understand. Know what your options are.
Can you walk me through your procedures for major incident handling, what happens in case of flood, fire or plagues of locusts?
If a physical machine fails and some guest VMs are migrated, no biggie. If power fails and UPS have to kick in, raised eyebrow but still nothing to be overly concerned about. If a major storm hits and basements all over your city, some of which are owned by your Cloud hosting partner are flooded – what then? Some of you will recall that this actually happened very recently. New York was hit by a major storm that resulted in flooding of hundreds of below-ground facilities of all types. Even the most hardened of facilities can suffer from incidents beyond control or consideration – what would your partner actually do in a situation like this? Don’t be satisfied by a double-sided A4 slick about “diesel generators, satellite backups and fire supression” ask to see the procedures for “zombie, locust or elephant attack” – these probably won’t exists, but you get my point.
Like most other decisions in IT, a large part of your decision to start moving to the Cloud is about understanding the risk associated to it.
Asking the right questions of partners and providers will help you balance the risk in a structured way, hopefully helping you to make the right choices helping you to get the very best out of the Cloud and what it can offer.
more to follow…