What is our primary use case?
Customers can use it for the web-based management of the product. We also store and retrieve data for their network connections. Also, we use the AI/ML portion called SageMaker to calibrate the algorithms and basically drive automation into the customer's use case. Typically our use cases are in hotels, public transportation, convention centers - anywhere where you are sharing internet connections. For example, hotels, conventions centers - anything where you might have people jockeying for a shared internet connection with possible oversubscription or network congestion. We also have enterprise Work-From-Home users due to the pandemic and they need to continue to provide access to those remotely into their own data center, corporate network, and public cloud.
How has it helped my organization?
Flexible fast way to bring up servers and network infrastructure with variable costs.
What is most valuable?
We use the AI/ML Sagemaker to help us build models.Â
We use several feature services on AWS, including Lambda, S3 database, RDS database, Alexa Voice Services and Cognito Gateway. They are all excellent in terms of offering great functionality.
They're pretty good about taking customer feedback and are generally able to productize the requested feature.
The initial setup is straightforward, especially if using Lightsail to start.
The solution scales very nicely.
The stability is good with a large number of Availability Zones WW.
Technical support is helpful and responsive but you must pay for a tiered support plan to ensure response.
What needs improvement?
The pricing is something you have to watch. You really have to constantly optimize your costs for instance, storage, IP's and things like that. That can become a job in itself to manage just from a budgeting standpoint if you are a moderate to heavy user. However, that's true for Azure or GCP as well.Â
If they did more automation on alerting you to cheaper pricing or automated volume pricing based on time/use or even porting you on to on-demand instances automatically, that would be kind of cool. That's something that I haven't seen yet. They could just automatically optimize for your workflow and put you onto a lower-priced instance to save you money. you Maybe allow you to pick an economy setting, or a performance setting, by time of day etc. something like that. That would be great. Then you don't have to think about it as much as you do in the current iteration.
It would be interesting to have a cost optimized accounting service so that they would come in and help remediate and give suggestions on how to cut costs. I know it's probably antithetical to their bottom line, but that said, obviously, if you take the high road there, you're going to probably keep people, and keep people from switching for lower costs. A lot of times, they can architect a better solution or a similar solution for lower cost and that would lead to customer retention--or maybe a longer term retention discount if youve stayed with them for awhile. That would be helpful if they had that. They have solutions architects, to consult however, they're usually just trying to design the best technical solution as opposed to the most cost-optimized solution.Â
For how long have I used the solution?
We've been using the solution for about four years at this point.
What do I think about the stability of the solution?
Services are pretty good stability-wise. They've got great redundancy. The one thing I would tweak them is when you're within the region or zone, they make it more difficult for you to do redundant zones, without carrying the IP addresses over seamlessly. That is a little bit of a sticking point, so you could have remote redundancy with the addressing there with it even outside of the AZ's. That would be a lot easier than having to go through the programming of it.Â
What do I think about the scalability of the solution?
The scalability is great. You can go from one small instance to GPU, very powerful instances, clusters. There is not any problem with scaling if you can afford it. If you've got the volume, you certainly can scale.
We have maybe a dozen or so customers that will use the product and then access the UI and the management system through the cloud. Then, of course, as developers, we have about 10 to 25 employees that have to use it to varying degrees to support the customers and do development.
How are customer service and technical support?
I like the tech support. It varies by level in that you've got to pay more to get the immediate response time. Generally, I'd say it's pretty good. Literally phone rings minutes after you log a trouble ticket. They're usually pretty good about escalations and helping. Out of AWS, Azure, and GCP, I'd give them the number two ranking. Azure has good support, however, it's expensive. GCP probably is number three I'd say, of the top three.
Which solution did I use previously and why did I switch?
We also occasionally use the Google Cloud Platform and Azure, although we tend to use AWS the most. GCP is a little bit cheaper overall, however, then you've got the cost of management that is typically a person so you do need to invest in that.Â
We started with Amazon and we've pretty much stayed with them. We've switched to Google and done some work on Azure that was customer driven, however, pretty much our prime public cloud has been AWS.
How was the initial setup?
The initial setup is not overly complex. It's pretty straightforward.Â
It's pretty easy to get started. However, you do have to make an investment and learn the different cloud platform's nomenclature. Most of our guys now are cloud practitioners and architects now that they've taken the training. We had to bite the bullet even though we've been users for four years. There is an investment that you have to make on the OPEX side. That's the case for any of the public clouds. Although once you know one, you can pretty much pick up the other ones pretty quickly.
What about the implementation team?
What's my experience with pricing, setup cost, and licensing?
Have to watch price/billing creep, but there are tools to watch and monitor your usage and billing. Â
Which other solutions did I evaluate?
What other advice do I have?
We're a software development group building specialty LAN/WAN optimization solutions, so we don't use a lot of canned products per se.
We do tend to sue reasonably new software versions of the OS...whatever is the latest LTS selections.
If you already have your workload ready, that's helpful, as you can actually trial it under a free tier and then see what the cost is, and extrapolate what the ongoing cost is. In the end, that's what gets you. Being able to do some benchmark testing on how much it's going to cost for your particular workflow across the three public clouds is definitely something you probably want to do. Especially if you're going to scale, as, obviously, it can suddenly creep up to not just tens or hundreds of dollars a month, but thousands a month, depending upon what you're doing. I definitely would recommend doing some reference testing of your workflows before deciding on a solution.
I'd rate the solution at an eight out of ten. They're pretty solid. You've got all the services that you can imagine, and then some. There's a very broad breadth of products and services. We haven't had too many SLA issues for recovery or downtime. Maybe we've just been lucky or good so far...
Which deployment model are you using for this solution?
Public Cloud
If public cloud, private cloud, or hybrid cloud, which cloud provider do you use?
Amazon Web Services (AWS)
Disclosure: I am a real user, and this review is based on my own experience and opinions.
