How feasible is this?

ShepsCrook

[H]F Junkie
Joined
Feb 16, 2003
Messages
9,439
So my company has a strong advocate for moving ALL operations to the cloud. I'm not a fan of the idea (at this time) as I feel that it would be extremely expensive and a poor decision until the company is more streamlined and operating smoothly in IT.

5 physical servers (file server, backup, 2x ESXi VM Hosts, DC)
13VMs (4 core CPU, 16GB allocated per)
30TB of Active data
23TB Used on SAN for VMs
50TB for backups
450TB of bandwidth (In and out)
200-250 employees (6 locations across the US)

Other cons that I feel make this a bad idea at this time:
Phones are non voip (Working on that)
Windows 7 systems still in use (Need to migrate most machines to 10)
Office 2013 across 98% of the company (licensing ran out and we are working on licensing for O365)
No imaging platform, patch management.
2 different domains (one for AD from old company, 1 for email with new company)
Company just invested in new 6 new switches for improving network
No solid onboarding or offboarding procedures
No solid policies for IT at this time (Passwords never expire, same password used to set up user accounts and email and never forced users to change password)
2nd ISP never connected after years of paying for it
I'm sure there's more but this is fine for now, I think you get the idea that IT needs an overhaul.


I mostly feel that the cost of moving to the cloud, the monthly cost would be to high, and with the rest of the stuff on the list needing attention, there's not a lot of time to deal with moving to the cloud. We'd just be moving some of those problems to the cloud with us and making it more difficult.
 
I'll keep this short since I am short on time. Managing in the cloud doesn't remove the management of your systems so much as it removes the need to keep up with licensing, etc. The group you work for needs to do some in-house maintenance before moving to the cloud unless they move to a fully managed MSP.
 
Moving to SaaS offerings is a good move for most companies. Often times it enables them to purchase a real enterprise-class solution for a fraction of what it would have cost them to build/support something on-prem.

However, moving your internal or customer-facing systems to the cloud shouldn't be a blanket statement. If you simply lift and shift your traditional legacy applications to the cloud, your management will be extremely unhappy with the cost and availability of the final result. In order to achieve the maximum benefits of cloud you will need to re-architect your applications to take advantage of things like autoscaling. In the cloud, you don't build for peak traffic, you build for min and let auto-scaling take care of matching your compute requirements to user activity. If you don't adopt this approach early, your management will get hit with a surprisingly large bill.

You will also need to decouple your applications to operate completely independently of other layers because things can and will die unexpectedly in the cloud. You are expected to architect for failure. If you don't your uptime will suffer.

You are right that your company will need to implement policies to control who can do what in the cloud. These are typically the same or similar to what is already in place.

I would support the move to cloud for any company but also caution against rushing in. It sounds like you have a pretty good understanding of what is going on with your company. Hopefully, you are in a position to create that task-list and help then work through it.

The biggest thing I would recommend (I can't say this loudly enough) is that a complete cost-analysis is completed for each application prior to the first click in the cloud. Also, work with your cloud provider to establish an Enterprise Agreement so you can get the best prices possible.

From a career perspective, cloud is the future of the IT industry. If you have a chance to get exposure, definitely do it.

BTW, I worked as a consultant at many of the worlds largest companies for several years. I can tell you from first-hand knowledge that no company is perfect. In fact, some of the largest companies have the most problems.
 
***Shortened***

I am in a position to fix things thankfully. Which is why I'm doing a lot of research on the subject and came here as well. Just trying to get the most information regarding this. My opinion and what I want to try to convey is while the cloud is great, I think at this time it's just a poor decision to try to just jump to the cloud. Build out a solid base to work from, clean up what you've got to be more streamlined, then move to the cloud if you become comfortable and confident with your setup. Right now, there's just to much wrong that needs to be fixed.

I do appreciate your information on this. I think our bill will be rather large and they will scoff at the potential $22-23k per month for "cloud" operations.
 
I worked on a project where we established backup to cloud as a substitute for tape and AWS with storage policies to move the older stuff to Glacier. If your looking at a way to dip your toes in the water and come off as the cost-savings hero, that's a pretty good first step.
 
You can leverage the Cloud Migration Initiative as a means to get a lot (all) of the changes you deem critical, as well. Think of "the cloud" as a potential green field to do a new deploy to. Start small. One org, or product, or app... or even small team moves first.

Most of your list doesn't matter if it's on-prem or on a cloud provider. Shit IT is everywhere, with poor and uneducated decisions coupled with "that's how we've always done it" being a common refrain.

Right-sizing your workloads and, as mentioned, aggressively using auto-scaling is definitely a large comonent of keeping cossts in line. The fact that all of your VMs are identically provisioned makes me knee-jerk react, thinking that none are properly sized.

Also, realize, being with a cloud provider allows you to move from CapEx to OpEx, which is desierable for some companies and orgs.
 
What can you automate? What needs to be on 24/7 vs what can be spun up on a schedule, can you tier any "cold" data to object(s3/blob/other)? etc etc etc. I have been down this road with a lift and shift and the price tags were enormous. The best way to approach is to catalog and score all applications that may live in the cloud and score them based on ease of movement to cloud, or cloud services, what apps and services have inherent dependencies on each other. And begin migrating applications based on that. This is a couple short sentences, but I could go on for hours on this topic.

I am a big fan of cloud, but if people haven't figured out how automate/orchestrate on premises systems, applications, lifting and shifting VMs and Data will be prohibitively expensive, then moving it back will be a PITA. There are products, 1st part and 3rd party for saving costs on data...eg, cloud.netapp.com and their ontap in cloud, CVS/CVO, Azure Netapp files on Azure, etc. You can take advantage of efficiencies on your data with these tools just like you could with your SAN/NAS on premises. Also have builtin in tiering to Object.

Bottom line, educate, educate, educate before stepping into this potential land mine. Cloud is not an easy button as much as they want you to think it is.
 
I used to architect and deploy VMware assets.
I moved on to AWS in 2015.

I was sitting at an AWS Loft a couple months back with 2 of the first 5 VCDXs.

VMware is partnered with AWS.

Bodies from Palo Alto, Juniper, and Cisco were there. SDN was out in full force.

Connecting resources isn't an issue, you either talk to your Enterprise rep and work out a deal on transfer costs for a hybrid stage or order a Snowball when you're ready to go all in.

Architecture and deployments are not a big deal, be scrupulously correct and public cloud is not different that datacenter assets.

We need to move on, and there will be a next thing down to road to learn.
 
So my company has a strong advocate for moving ALL operations to the cloud. I'm not a fan of the idea (at this time) as I feel that it would be extremely expensive and a poor decision until the company is more streamlined and operating smoothly in IT.

5 physical servers (file server, backup, 2x ESXi VM Hosts, DC)
13VMs (4 core CPU, 16GB allocated per)
30TB of Active data
23TB Used on SAN for VMs
50TB for backups
450TB of bandwidth (In and out)
200-250 employees (6 locations across the US)

Other cons that I feel make this a bad idea at this time:
Phones are non voip (Working on that)
Windows 7 systems still in use (Need to migrate most machines to 10)
Office 2013 across 98% of the company (licensing ran out and we are working on licensing for O365)
No imaging platform, patch management.
2 different domains (one for AD from old company, 1 for email with new company)
Company just invested in new 6 new switches for improving network
No solid onboarding or offboarding procedures
No solid policies for IT at this time (Passwords never expire, same password used to set up user accounts and email and never forced users to change password)
2nd ISP never connected after years of paying for it
I'm sure there's more but this is fine for now, I think you get the idea that IT needs an overhaul.


I mostly feel that the cost of moving to the cloud, the monthly cost would be to high, and with the rest of the stuff on the list needing attention, there's not a lot of time to deal with moving to the cloud. We'd just be moving some of those problems to the cloud with us and making it more difficult.

The only thing in your post that I see that should really affect your cloud move are these:

No solid policies for IT at this time (Passwords never expire, same password used to set up user accounts and email and never forced users to change password)
Cost

The other items aren't really relevant to a cloud migration unless you are saying that its a lot of extra work you have to do. But it didn't seem that way since you pointed out investments in switches etc. Those aren't going to go to waste.

A cloud migration when things are that bad can actually improve things because you get a fresh start. If its really that fucked up maybe its better to just start with a clean slate? You'll have less resistance to change - for example implementing a password policy and 2FA because "well its in the cloud and this is part of that move..."

Have you run a TCO calculator on cloud vs on prem? Also exactly what would you be moving? Just servers to VMs hosted in AWS (hint: that's not a cloud migration that's a hosting change)? Or would you shift to use SaaS?
 
Back
Top