Azure Stack's debut ends the easy ride for AWS, VMware and hyperconverged boxen Microsoft's formal announcement of Azure Stack changes the cloud market forever, by giving us three distinct types of cloud. Today we have pure-play public cloud from AWS, Microsoft, Google, IBM or what I've heard described as “hybrid virtualization” – the ability to stretch an on-premises private cloud into the public cloud …
I disagree - integrating with existing systems is the opposite of the point of this. It's not an interface, its a different way of working. If you want to spend months working on integration then buy traditional stuff. If you want hassle free deployment that just works buy this. crowbarring your existing stuff into this will make this stop "just working" so why put yourself through it. Just let the old stuff die off gradually or do a migration and start realising the TCO savings and value model.
This is all assuming HPE, Dell etc. can actually write some stable drivers, of course, and that's a big ask as we've seen with VMware and Hyper-V on prem over the last decade.
Do you understand what this is? I'll guess not, but grats on first post before even performing the slightest research.
This is a cloud platform. It is not about launching 1980's era tech on the latest and greatest systems to manage. It's about giving developers a platform to write applications for which can be hosted in multiple places without the interference of IT people. It's about having an app store type environment for delivering systems that scale from a few users to a few million users. It's about delivering a standard platform with standard installers so there is no need for building virtualized 80's style PCs that require IT people running around like idiots to maintain.
You can keep your VMs and SANs and switches. There are a lot of us who are already coding for this platform and can't wait for this to fly. Whether you like it or not, we're going to write software for this. Your boss will buy the software and either you can run it or you can find a new job :)
I think eventually this will be great. At launch there are just about enough services to convince people to get on board assuming that live migration is included (an issue in Azure) for maintenance. I'd certainly buy it for IaaS over VMware any day. Once all of the PaaS stack is ported, and that shouldn't be long, this will be a truly great platform and a big enabler for development.
Personally I can't see a world where AWS doesn't respond. All of their stack is based on open source and they have hundreds of genius coders working to a very effective DevOps model. The only thing holding them back is a pricing model, and they are certainly smart enough to cope with that. It's possible they could rent hardware, or resell own brand hardware, they already have snowball and snowmobile so they aren't afraid of hardware. They already have SPLA licences in place so MS licensing wouldn't concern them either.
This is all very good for consumers, and it's nice to see a proper tech fight again which pushes things forwards. That said, unless there's a good reason to be on premises such as being on a ship out to sea I'd still go with real cloud. After all, who wants to be swapping out failed drives and fans these days? Even compliance has caught up with public cloud in most sectors.
There's absolutely no reason for AWS to enter this space. It's probably a sizeable market, but Amazon know where the growth lies and what they do best. Putting kit on premises makes no sense, except where it represents an on-ramp to public cloud. It's not that long ago that Microsoft promised to ship an Azure software stack that would effectively compete with OpenStack. Now, they have now decided to ship the appliance that they promised they'd deliver back in 2010. As I say, there's a market for this, but it's hard not to see it as largely transitional.
I have customers spread out across government, finance, medical and even publishing that can not use public cloud because of safe harbour. They all want cloud, not virtual machines, but end to end PaaS and SaaS, but they couldn't have it because one way or another, you're violating data safe harbour laws or simply giving your data to China or India. This is a huge thing.
I've got this issue with my clients too, but I know for a fact that all of those who have safe harbour / data protection issues still send a lot of sensitive data around the world (think spreadsheets etc...) in unencrypted emails, and make use of sharing services such as Dropbox and GDrive for work related material. I know they shouldn't and you can tell them they shouldn't, but they still will. On premises hardware solves none of that.
"Personally I can't see a world where AWS doesn't respond."
Which pile of freeware crap do they adopt? Or do they write their own? What about all the tool sets that need to go with it for management, monitoring, etc, etc. (that Microsoft already have) ?
Microsoft are just so far ahead in on premise that AWS don't have a hope in matching this in any reasonable timeframe imo. Everyone I speak to thinks that Azure is eventually going to be market leader in cloud because of stuff like this...
When Microsoft first announced Azure Stack, they were open to customers running it on their own hardware and the test systems seemed to indicate that this was possible. I guess they've rethought their ability to support totally random hardware configurations, and are now only supporting Stack on "integrated appliance" type hardware. The problem is that I haven't seen any of these appliances from vendors that scale down further than a rack-size footprint. The idea here is that you replace a good chunk of your on-premises hardware with the Stack, pay Microsoft and the vendor as you go, and the vendor keeps your appliance fed and happy IBM mainframe style. What remains to be seen is how much both companies will charge you to run stuff...obviously the PAYG rates will be cheaper since it's your hardware, but I can see stuff like PaaS SQL and other toys getting very expensive to run long-term.
If vendors are willing to put out a "tiny" version of this, I could definitely see companies with lots of small locations (like us) that can't put stuff in the cloud buying them. So far I haven't seen any yet.
The entry point is 4 nodes and then scaling in 4 node blocks to 12 nodes in a "Region". (Scaling not initially available, ETA 9 months)
You can deploy multiple "Regions" in a single SP or Enterprise customer and migrate between them as you would Azure regions, also replication to Public Azure is supported.
"wtf is OpenStack then"
A horrible to use VM management / deployment system (editing text files to store config in the 21st century?!) that doesn't easily plug into any major commercial cloud provider like Azure stack does into Azure.
On the plus side though, if you have to torture yourself with OpenStack - it works well with Hyper-V - which provides the fastest performance under OpenStack of common hypervisor options as per recent benchmarks!
> Seriously folks, Amazon is the Microsoft of the 21st Century. They are an existential threat to everyone who is not them. Avoid them at all costs.
Except that Amazon have great products, competitive prices, high uptime, and surprisingly good customer service; all things which aren't Microsoft.
If buy into the full AWS stack from Lambda via SQS to DynamoDB, then you certainly have your hands tied vendor-wise. OTOH, those who've done this tell me it works really well.
This is a brilliant play by Microsoft. In many heterogenous networks Windows is slowly being relegated to a middleware software layer running on the likes of VMware and being accessed from thin clients and mobile devices. This turns the tables right around and puts their stack out in front, with Linux and other platforms becoming the meat in the Microsoft sandwich.
The trick for those wanting to go down this road will be to watch the early adopters and then jump in once (or if) this matures - execution is certainly not Microsoft's strong point historically so best leave to others to sort the teething issues out first (and there will be plenty). If this works out it will make for a very low entry barrier for those who simply can't use public cloud.
Like most 'black-box' solutions, the drawback is massive potential for lock-in via proprietary APIs so it will be interesting to see how this plays out long term. It might be an ideal solution medium turn, but your entire infrastructure becomes dependent on the direction of a single company which always results in pain when you have a business need that doesn't fit into the mould.
So why do developers go to the cloud? Because its quick, no internal lengthy capex delays, easy to scale fast and predictable costs (on the most part). This is just a converged box with Microsoft Azure on it - it will sell, a little but in the end....tin WILL die for 95% of applications. Everything will be delivered via browsers, so windows will die, serverless compute will diminish server licenses, SQL will still be strong for a while but who knows.
I see Microsoft playing the Hyper-V game with Azure. They are always looking at AWS and copying in the later release, as they did with VM-WARE. Google still has a strong part to play with "the cloud Wars" due to their far superior analytics and search engine but I can genuinely see, unless they start to innovate rather than copy, Microsoft becoming an IBM or HP...dead horses
To be fair to them, AWS have been copying some features the other way recently. Since MS finally got on the Agile bandwagon and started delivering they've really been pushing forwards at pace. They are still behind in some areas, but others like Logic Apps and BizTalk services were firsts on the Azure platform with AWS playing catchup.
But yes, the photocopiers nearly melted over the last three years or so as they caught up to Amazon.
Google don't have the scale or experience right now, and don't even seem to have the same clarity of vision as AWS and Azure. I really hope they start to catch up as a two horse race is never ideal but right now that looks to be the case. Google certainly don't have superior analytics, they do have fairly good marketing but Amazon and MS, and IBM in that space are all very competetive. I agree HPE are not where they need to be, but then neither is Oracle and I suspect they will all be around for years yet while their IP and money last.
"After all, who wants to be swapping out failed drives and fans these days? Even compliance has caught up with public cloud in most sectors."
..........Of course you don't, but somewhere, someone will also be doing this in the "Cloud". Zero impact on customer services, via clever QoS delivery models? Maybe...
It' s as if the whole IT world has become so blinkered by notions of abstracting apps, DB, n code etc away from the nasty HW layers, that they have completely forgotten the underlying Cloud Tin, which is still as frail/robust/prone to inter-tin/firmware glitches as it ever was.
"Ah", they say, "But MSFT, AWS. Google etc can all pick that up, absorb the massive interopereabiliy and microcode revisions testing bills, cost-effectively, whilst giving everyone their desired OLTP, uptime, RPO and RTOs."
As a CTO/CFO/CEO, are you going to bet the Company's mission critical apps and data on that? Performance and cost are one thing with Public Cloud, data sovereignty, security, and latency are another. One can see why the big ISPs are pushing for abolition of net-neutrality.
Many CEOS will still be requiring their teams to continue swap fans n HDDs for some time to come, I suspect.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
