Not a service; the company.
Not a service; the company.
At least go with Dell. Dell will sell you an MD3860i with 60 6TB hard drives for not much more than what you paid for the your Synology. Performance is just as good as the Synology, you'll get next day on site support from a Dell tech, and a smaller rack/power/cooling footprint as well.
Seconded... though having recently seen a lot of quotes you could do worse than the Dell SCv2000 which is the newer replacement for the MD3860i using the Compellent code. It's faster and cheaper than the MD, mostly because Dell no longer has to pay the Netapp tax for every MD (the MD's are based on an LSI chipset that's owned by Netapp)
There's a reason nobody builds deskside compute servers with today's technology. Density, power, and cooling.
And the fact that a deskside system is highly unlikely to be utilized 100% of the time... probably more like 10% of the time. In that case it's more cost-effective to farm it out to a bigger cluster in a server room, or run it in some AWS/Azure nodes for the time it needs and then shut it down.
The fact that there are many more high performance computing resources available relatively cheaply is as good a reason as any not to do deskside compute on a large scale.
I wish you hadn't posted AC, and I wish I had mod points!
This is the right answer. The workloads aren't clear because we don't know what OP is trying to accomplish with this setup. Is he building an HPC cluster to do engineering analysis, or is he building it because he has convinced his management that it's cool? If the latter... well, he'll be looking for a new job after he builds it and it does nothing to help his customer (his employer).
Start with the application. What are its workload characteristics? What kind of backend does it need?
Then look at the backend. What are ITS workload characteristics.
Only then should you look at hardware. Answers to the first two will answer what hardware you need, but even then there are a lot of moving parts to take into account. How many nodes? How much CPU per node? Highly parallel or high speed? How much memory per node? Storage infrastructure? Output type?
This is a terrible AskSlashdot question because it requires an intimate knowledge of the workload being proposed. I am a consultant for a living so I do exactly this process above every single day. Any answers will literally just be spitballing because the information we have available is so vague... and any actual answers are guaranteed virtually useless and a quick way for OP to lose his job.
I know this is old now, but honestly you're overthinking this.
First, as others have mentioned here you can use TeamViewer to do remote desktop support, and it's free. No need to upgrade to Windows 7 Ultimate or anything else for that matter. I've used it on OSX, Windows and Linux and it works like a champ. I've supported family and friends... and even had a commercial license for TeamViewer for a while because it really is so easy to use and maintain that I found it invaluable. I don't do that job any more, but I still maintain TeamViewer on my computers in my house so I can get into them and manage/maintain them while I'm on a business trip. Same on my son's laptop so if he has a problem I can support him remotely.
Now of course comes to your son. Don't. Seriously... kids are going to be kids, and they're going to work around any controls you put on a computer. The only thing you are LEGALLY required to do is to control what he has access to at YOUR home. Once he's off your network, anything he does is the responsibility of the party that owns the network he's using. Yes, he should be held responsible by you as a parent, but legally there's nothing forcing you to do this. Plus, kids are going to find workarounds regardless; my son is 15 so you can imagine the battles I've had with him over the years. As it stands now, I manage his Internet access at home using a Sonicwall TZ-215 firewall that has Gateway Anti-Virus and some content controls turned on. Honestly, I don't block porn... he's 15... but I do block some categories I personally find distasteful; hate speech and the like. If he needs something for a particular essay he's doing for school that's blocked, he can ask me to unblock it and he does. This way there's mutual trust going on, which to be honest is the RIGHT way to parent.
I also don't check the logs to see where he's going on the web. Just so long as he's not doing anything illegal (and yes, I do block bittorrent for that reason) that could get me in legal hot water I don't particularly concern myself with it. I check his laptop for malware and to make sure updates are in place periodically, but beyond that I don't see the need to get overly stressed about it. Besides, we have an understanding that if he does anything bad that gets his computer malware that's going to be too much trouble to clean up (like more than 30 minutes of work on my part) then his machine gets re-imaged and he gets to reinstall everything, restore his own files etc. I make him responsible for his backups as well.
Is my system perfect? No, but it works. And right now I have a 15 year old boy who may or may not go on porn sites occasionally (I really don't care), plays games occasionally... but generally is a well-behaved kid when it comes to technology.
I guess what I don't get about your requirements; if your primary reason for the site B connection is supporting your parents, then why backhaul all the Internet traffic across your own network? With a decent managed firewall you can do all the controls you like, and there are web-managed options as well. Some of them even support OpenVPN natively or some IPSec variant that you can create a virtual private network for managing stuff. If you really want content controls on your parents network then you really need to review what you're trying to accomplish here. You don't have to get something as fancy as a Sonicwall, there are plenty of other cheaper options but that is certainly one.
I do have a VPN as well as my TeamViewer connections... honestly SSH is easier to manage my Linux boxes than TeamViewer most of the time because I don't need a GUI. As a result, all my Linux boxes partake in an OpenVPN network against a hub system hosted on Linode (where my web server is also hosted). I have the OpenVPN client on my laptops so when I'm out and about I can join the network and SSH to any of the systems no matter where I am (I keep a HOSTS file with all the IP's). Bonus; I can host my own mail server on my home box without using the storage on the Linode because I have it stood up as a VM on my home server... the Linode has a pretty basic Postfix configuration that relays to the private interface via OpenVPN, and the mail server is configured to send mail out via this same relay method. If my mail server goes down then said mail just spools on the Linode until it comes back up. Since I only have a handful of users (family and friends) it's not a big deal. For the record, I use Zimbra as my mail server for simplicity's sake
And yes, the Sonicwall also partakes in the OpenVPN... so if I connect to my VPN from an hotel in another city, I can then browse to my Sonicwall's web interface and manage it. For me it's better than "cloud managed"
I hate the fact that you were modded as a troll, because you're completely right. I use MobaXTerm on my Windows 8.1 box and it is great for managing all my Linux boxes.
I spent years trying to get a really functional Linux desktop, but now I'm on Windows and quite happy. Yes, I still use Linux for servers but generally for desktops I see little point in "rocking the boat" so to speak. There's just too much hassle to get Linux happy on some modern hardware; I still have to custom compile a kernel in order to support the WiFi card in my laptop purchased in late January 2015... the appropriate drivers are only available in the bleeding edge kernels in bleeding edge distributions... and to be honest I'd rather just have a laptop that works than one I have to futz with periodically because something broke.
Yeah, I'm getting conservative in my old age... or maybe I just want to get work done.
To expand on this, it also depends a lot on the job. While you're absolutely correct for software developer positions, there are other completely different positions that offer work from home that actually works.
One prime example is my job. I'm a consultant who works for a (very) large technology company. In my role I have very clearly defined deliverables that require me to get off my butt and do stuff; namely customer visits, presentation, system designs and so forth. And my pay is structured such that I can live on my base without lifting a finger, but I will live a lot better with the commissions I am paid in addition to my base. Yes, my role is partly sales but I have found I am quite good at it (which was a surprise for someone who spent the last 15 years or so locked up in a datacenter). This means I am measured but also driven... and I get to see an almost immediate return on my investment of time and work.
My job is one that works particularly well for work from home. If I don't work, I get paid less... and eventually those "measurables" show that I'm not doing anything. It might take a couple of quarters but eventually I'll get replaced and that'd be my own fault. But if I work hard and do the job I'm asked to do then I see improvements in my paycheck that encourage me to work even harder, and the statistics show that I'm doing my job. Everyone wins, right? Yeah, there are catches but generally it's good for everyone involved. It doesn't hurt that I enjoy the hell out of my job.
Obviously this doesn't work well for all jobs. My job requires customer face-time so it can't be sent to India. Though theoretically about 80% of my job probably could be done by someone in an office in Bangalore, it's that 20% "soft-skill" stuff that can't. 99% of my customers are quite conservative and would NEVER accept someone trying to do my job via Skype or some other technology. It just doesn't work. By the time the generation of kids coming into the workplace who are weaned on Facetime are in the CIO/CEO/CFO positions that I typically talk to I'll either be dead or retired. Software developers are particularly vulnerable to this problem because they are doing work that can be easily offshored. This is why WFH doesn't work for every job.
I guess my point of this ramble is that Work From Home is possible, and can be extremely enjoyable (I consider it a great perk of my job) but first you need to have the right kind of job. And if you don't want to be constantly concerned about losing your job to someone in India or China then you need a job that's customer-facing and profit-making for the company. Software developer is a no... customer-facing consultant is a yes. Finally, a good incentive program to work hard is not required but certainly makes it a lot easier. Just be wary of the risks of such an endeavour; if your boss doesn't also work from home then you might find yourself excluded by the fact that he doesn't see you every day. Mine works from home in another state so I rarely see him... but he covers an entire region so he rarely sees any of his people in person. But again there's that mindset; he works from home too so he knows how to work with employees who also WFH.
If the system is architected well, shouldn't all of those steps be automated... including monitoring and failover success/failure?
In a perfect world, with perfect systems documentation you'd be right. Unfortunately few of us have the pleasure of working in such an environment
I had an SP1 and now own a Dell Venue 11 Pro (competitor to the Surface/Pro 3). What exactly is the problem with it? I realize it depends a lot on the applications you run, but that is true of any device. For example, if you're using it to run Putty... then no, the SP or its competitors are going to suck. However, if you're using it as a content consumption device, which let's be honest is exactly what a tablet is made for and good at... it's awesome. Now, that comes with the caveat that you're either using apps or the touch browser... desktop side is different.
But on that desktop side, you can run Putty... you can run PowerShell... and you can run any one of a number of desktop browsers and applications, but of course then you have the OPTIONAL keyboard that you can use to operate said applications.
The beauty of the SP3 and its ilk is that it is exactly what you need when you need it. No, none of them are going to be incredible gaming rigs but they're awesome at being productivity tools as well as content consumption devices. If you need solely or primiarly a productivity tool then no the SP and its ilk are probably a bad match. However, if you want a tool that can do both jobs equally well then you really have no choice but to go with one of these Windows 8 devices. But that's not a bad thing, either.
For $500 you can get a laptop that's built like a Soviet tank, or has atrocious battery life, or has a really slow SSD, or has 2GB of RAM... you get the idea. For someone who wants performance, portability, storage and being useful there aren't a lot of "magical $500 laptops" that really fit the bill. I know; I went shopping recently for new computers (yes, plural).
I ended up with a Venue 11 Pro and an Alienware 15. Yeah, both Dell but they fit my needs perfectly. The AW is my gaming rig and is awesome (runs AC:Unity in Ultra and looks gorgeous) while the VP11 is my computer of choice when I'm on the road (I am extremely mobile at work, traveling sometimes hundreds of miles in a day). Both of them are attached to the same Microsoft account, have Evernote, OneDrive and DropBox installed. As a result, the same files are always available and my workflow benefits greatly from it. As files fall into disuse I migrate them off to the "archive" which is my Linux-based server at home... which is still accessible via an OpenVPN that I have installed on all of my computers if I happen to be on the road.
But you know what? Neither of them will do "all the things" perfectly. They both run the same OS and the same applications, but serve completely different purposes. I would never sit in front of a customer with my AW in part because the battery life is horrendous (it's a gaming laptop, what do you expect?) and is just not really all that portable. Similarly I would never consider playing AC:Unity on my VP11 because... well, it would suck quite frankly. Could I? Yes... but just because we can do a thing does not follow that we should.
So why did I buy two laptops instead of a desktop gaming rig? Well, that was so that when I am traveling for several days (have a trip to Chicago coming up in a few weeks for example) I can still bring my gaming rig with me so I can play in a hotel room. Sure, it rarely leaves my desk but when it does I want to have that powerhouse on-hand.
Well, I don't have a Surface 3 as I upgraded my Surface Pro 1 to a Dell Venue 11 Pro (7140; late '14 model). The same Geekbench test turns up a number of 4364. That's running the Core M 5Y71. I'd expect the Surface 3 to be similar in performance. The x7 is slightly faster than this Core M but doesn't support dynamic frequency scaling. Not sure what ultimate difference this would make to battery life.
Having said all this, I have no complaints about my Venue. It works exactly as advertised, is reasonably fast, and has an absolutely crazy-good battery life.
So you were a OS X dev for years until recently yet you think that "Driver lever changes" can bring OS X's power management to Windows? You don't even know what Grand Central Dispatch is but you were an OS X Dev? Serious credibility problems there. You're a troll. Ta, ta
Wow, you're hilarious.
I know exactly what GCD is, and have worked with it. I just am under this strange belief that there's no magic in software and that computers still succumb to the laws of physics. Better power management comes primarily from hardware... and while GCD is a very good framework for controlling that hardware it's not a magical route to better battery life. Software can help with coordination but you still need the hardware.
While I won't argue the concept and implementation of this framework are good, most of the magical gains you'll hear about because of GCD are marketing numbers and not real-world.
Yet other laptops running Windows get just fine battery life with the same or smaller battery. Simply put; Apple develops their own EFI to manage fans and power states and when the driver doesn't exist or is non-optimal, the hardware runs "hot". Grand Central Dispatch is a nice technology but does NOT magically increase your battery life despite what Apple wants to claim. It merely provides a common framework for addressing the EFI and power management customizations in a way that is pretty nice, but not a magic bullet. As Scotty was fond of saying, "You cannae change the laws of physics, Captain".
The level of ignorance and bias you display are unfortunately all too common. I get it; you're an Apple fan. Just because you read an article on a nice API doesn't mean you know diddly about hardware and power management. I don't claim to be an expert, but after doing embedded systems development for a while in my youth I think I might have some more knowledge of which I speak.
There is no likelihood man can ever tap the power of the atom. -- Robert Millikan, Nobel Prize in Physics, 1923