5 ms·
Slightly nitpicky, but is something a PaaS if you run it yourself? Anything as a Service isn't a service if you're running it yourself. It's just ... infrastruc
by onion2k 6y ago
Slightly nitpicky, but is something a PaaS if you run it yourself? Anything as a Service isn't a service if you're running it yourself. It's just ... infrastructure.
- ashnehete 6y agoI was just thinking the same thing - maybe we can call it self-hosted PaaS.
- baq 6y agomaybe you want to decouple sysadmins from devops? sufficiently large orgs with sufficiently large on-prem infra have these kinds of problems.
- idoby 6y agoI just read it as short for "cloud infra project with controls modeled after popular PaaS products"
- vincent_s 6y agoYeah, it's more like a PaaS alternative
- bomdo 6y agoNot nitpicky at all - this is an important distinction to highlight for pointy-haired decision makers. This product is undoubtedly the P in PaaS, but there is no service behind it. If your company uses this as an alternative to a real Heroku/AWS/xyz PaaS, you must have engineers at hand for 24/7 ops, scaling servers and fixing bugs. In my opinion, this is quite risky for anything running in production and should not survive a cost-benefit analysis.
- 1337shadow 6y ago> should not survive a cost-benefit analysis I completely disagree, the difference of price between dedicated servers and even EC2 instances is completely amazing. This is what you get for less than $200/month with a dedicated server: 1× AMD EPYC 7281 CPU - 16C/32T - 2.1 GHz, 2 × 1 To NVMe, 96 Go DDR4 ECC, unmetered 750 Mbps In one of my companies the AWS bill is just completely insane, we have like half that hardware, with a really small bandwidth, which is metered, for more than $800/month, which is fine while we're on free credits. I love working for cloud companies, it's a lot of fun, but when it comes to my money then I never go for anything but a dedicated server.
- zeveb 6y agoYes, hardware as a service will always be much more expensive than hardware you own. But it may be less expensive than the team you will require to run that hardware at an acceptable service level. It very likely will be less expensive than the opportunity cost of running your own hardware. As an example of the latter bit, if you are running your own hardware and need to add another host and you do not have a spare lying around, then you need to order one. It has to be shipped. Someone has to unpack it. Someone has to make sure that the data centre has sufficient power. Someone has to install it, its power and its network cables. Each of these steps takes time, but also each step is an opportunity for friction. By contrast, with a service, you would just add a new host. Five minutes later you are up and running. That gives you an operational nimbleness that you wouldn't otherwise have had.
- ricardobeat 6y agoThat example is not realistic. You rent dedicated servers from a provider that will always have extra hardware at hand, and handle all of those steps; you don't rent hardware yourself and run it in your basement :)
- zeveb 6y agoWhat you just described is a kind of hardware service!
- vidarh 6y agoOr you rent managed servers or colo space from one of the many hosting providers that also offers cloud services, and pick and choose. That lets you migrate your base load to colo or managed servers over time, while you still have the nimbleness of being able to scale up and down dynamically if you want or need to. And my experience from providing devops services to clients on a contract basis is that the clients who use cloud services tends to need more, not less, devops assistance.
- parliament32 6y agoI love how there's this myth that servers and services just blow up every 10 minutes 24/7 and unless you have a legion of ops personnel you're going to get hours of downtime each year. Servers, for the most part, just work. In DC climate-controlled environments, hardware failures is exceedingly rare. Apart from harddrives, most hardware will happily tick along for a decade, if not longer. Sane production-grade OSes (read: not Ubuntu) will also happily run for literal years with zero human intervention. For obvious reasons, it's a bad idea to not patch your systems, but things will continue to "just work" pretty much forever unless you're running really shitty code. For renting vs buying servers, there's upsides and downsides. Buying gear is far far cheaper if you plan to be around for more than a year, but renting dedicated servers gives you a lot more flexibility -- to provision a new server, you hit a button in their online panel, wait 15 minutes, then let your deployment strategy take care of the rest. I find it almost mind-boggling that AWS and friends have convinced people that it's normal to spend ridiculous amounts of money for fairly "meh" service specs in what's essentially VMs.
- 1337shadow 6y agoNope, you have IaaS (Infrastructure as a Service) like OpenStack, and PaaS (Platform as a Service) like OpenShift, KelProject, CapRover ... The idea is that you ask OpenStack a VM and it will give it to you, dealing with the lower level details for you. PaaS means that you ask it to deploy a service and it will deploy it for you, dealing with lower level details for you.
- ghaff 6y agoNeither you nor the parent are wrong but I'd argue that you don't really see IaaS or PaaS used all that much for on-prem platforms these days. (And the definition of PaaS was always a bit fuzzy--something like an abstraction that is in between IaaS and SaaS. You're probably more likely to see OpenStack called a private cloud or on-prem cloud than "IaaS" these days. And OpenShift is usually called a Container Platform rather than a PaaS.
- 1337shadow 6y agoThe definitions have always been pretty clear to me, but all right then, thanks for the heads up, I guess CapRover people and I are also what we call "old school devops" these days. "Container platform" seems pretty vague to me, PaaS means something I know right away. I mean, k8s is a container platform too isn't it ? But you'll need to build what we called a PaaS on top of it yourself (or use something like Kelproject, OpenShift ...)
- ghaff 6y agoYeah, the terminology isn't always super-clear. Yes, k8s is a container platform. OpenShift, depending upon how you use it, can span a range from being an integrated k8s distribution to something a lot more like what was commonly called a PaaS with developer tools, CI/CD pipeline, registry, etc. PaaS isn't a verboten term or anything like that. But it turns some people off because it was most associated with services/products/projects that mostly focused on a simplified developer experience at the cost of flexibility.
- cbetti 6y agoThe distinction you are drawing sounds to me like the difference between "managed" and "unmanaged." My read on whether something is a service or not is, can I make a request of the thing in simple terms, and have the thing carry out all the messy details on my behalf?
- jordanbeiber 6y ago”Service” to me is just ”delivery of something with a specific scope and a defined contract”, not so much about who delivers said ”service”. Many IT depts would do themselves a massive favor to deliver actual services instead of “just infra and some stuff thrown on top” and call it service delivery. Tools like in this link can help, but a big part is simply about automation and delegation/self provisioning.
- dubcanada 6y agoYou can't just decide that words mean different things to "you". Platform as a Service or anything "as a Service" means someone else provides it as a service (ie subscription). The Platform part is all this is offering. So it is not a Platform as a Service.
- jordanbeiber 6y agoYes, ”someone else” - be it delivered internally from one team to others or b2b. It is not necessarily hard tied to a business model, but of course I understand that this is the common usage. It’s really about abstractions and consumability. This is my interpretation of the NIST meaning of aaS.
- c17r 6y agoPaaSS - Platform as a Self Service