Jump to content

Recommended Posts

Posted (edited)
Not according to IBM.

 

"A SAN Volume Controller clustered system has several features that can be used to deploy a high-availability storage system with no single point of failure."

 

The SAN is specifically designed to avoid a single point of failure. Multiple disks, multiple network ports, multiple controllers, multiple power sources. All designed so that if one piece of hardware fails there is a layer of redundancy.

 

Minus the natural disaster damage of course!

 

 

That is specifically regarding IBM SVC not SANs which is what my comment was about.

 

You have your storage in one box?

 

How much was the original setup and how much for a complete hardware refresh? What was/is the benefit over a properly redundant and HA DAS server setup, which would be a fraction of the cost of a truly redundant and HA SAN setup?

Edited by JackCPickup
Double quote
Posted
That is specifically regarding IBM SVC not SANs which is what my comment was about.

 

You have your storage in one box?

 

How much was the original setup and how much for a complete hardware refresh? What was/is the benefit over a properly redundant and HA DAS server setup, which would be a fraction of the cost of a truly redundant and HA SAN setup?

 

We actually have 3 SANs. The original IBM SAN is 10 years old now and still working well. We could easily do without it now but as it works perfectly without missing a beat we can easily find a use for it, even for something low on the priority list. We still buy in SAN support for it so all parts are covered anyway for a nominal fee, £250 per year.

 

Our storage is split and replicated between 2 file server VMs on separate hosts using DFS. I'm pretty happy with the level of protection we have.

 

To be honest I think it would be poor management on my part if we were ever in a position to have to do a complete hardware refresh. The boss would go bezerk if I went to him one year and asked to replace a load of things in one go. We have a rolling programme so we know what we will be adding over the next 3 - 5 years. As we scale up we just add what we need and decommission the oldest bit of kit as it is coming to end of life. Managing the budget accordingly has allowed us to plan carefully all of the large expenditures for the next 5 years.

 

Everyone has their own way of working and it's not one size fits all. My attitude to it is that we are talking about the foundation of the whole network, so I'd rather not do it on the cheap if the budget allows me to put in the best kit. If you have an enterprise sized network then really I think you should be using enterprise quality hardware.

 

All this obviously depends on budget, size of network, control over resources, loads of things.

Posted

My question is. What does a SAN provide that a modern much less expensive 12GB SAS DAS does not? The only thing a SAN provides that I can see is the ability to provide storage access to more than than 8 VM hosts if you use a huge fibre channel switch.

 

How many schools are running more than 8 VM hosts? How many are running more than 4? So why SAN? Why do you need the N for "Network"?

Posted
My question is. What does a SAN provide that a modern much less expensive 12GB SAS DAS does not? The only thing a SAN provides that I can see is the ability to provide storage access to more than than 8 VM hosts if you use a huge fibre channel switch.

 

How many schools are running more than 8 VM hosts? How many are running more than 4? So why SAN? Why do you need the N for "Network"?

 

The fact that you can't upscale the SAS without doing the whole disk array is an issue. Performance is much better on the SAN, heck, it's even better on a modern NAS than a SAS, especially if it is your datastore for VMs and you need high throughput.

 

I can understand your logic , but the only reason I'd go with an inferior device is if that's all we could afford. Why would a millionaire buy a Bentley if they could buy a brand new Ford Mondeo at a quarter of the price? They can both do the same job, just to different standards. It's a fact of life that you get what you pay for.

 

For the most important piece of hardware on our network I would rather spend a bit extra. We have been running SANs for 10 years and my experience with them has been first class.

 

Our IBM SAN was £7k 10 years ago. £700 per year is pretty good value if you ask me.

Posted
The fact that you can't upscale the SAS without doing the whole disk array is an issue. Performance is much better on the SAN, heck, it's even better on a modern NAS than a SAS, especially if it is your datastore for VMs and you need high throughput.

 

I'd put a modern 12gig sas system against any SAN, performance wise.

 

 

For the most important piece of hardware on our network I would rather spend a bit extra. We have been running SANs for 10 years and my experience with them has been first class.

Our IBM SAN was £7k 10 years ago. £700 per year is pretty good value if you ask me.

 

But it's not £700 per year is it, because you've had to buy another two SANS on top of original IBM SAN? Why? I thought you said a SAN can expand and grow with you so you should have just kept on expanding your 10 year old one shouldn't you?

 

Just curious! :-)

Posted
I'd put a modern 12gig sas system against any SAN, performance wise.

 

Benchmark tests in various reputable hardware and IT organisations wouldn't agree with you there.

 

 

 

But it's not £700 per year is it, because you've had to buy another two SANS on top of original IBM SAN? Why? I thought you said a SAN can expand and grow with you so you should have just kept on expanding your 10 year old one shouldn't you?

 

We haven't 'had' to buy them. We could easily have upscaled the original device (something you can't do with a SAS) and kept them but it wouldn't be good practice to keep a 10 year old device as the main datastore when we can upgrade the system periodically with allocated funds. It's better to add technology into the network gradually rather than waiting for it to die and then replace it, minimising downtime and streamlining the installation of new equipment.

 

Any network grows exponentially with demand, users, devices, technology, so if we didn't buy the SANs we would have had to have bought something else. Storage demand has exploded as the internet has improved and the demand for video editing and stuff like that. A network of 10 years ago is very different to one of 2017.

Posted

Why is your DAS limited to 8 VMs?

 

Main reason to use network storage is for access from multiple physical servers, so the VMs can be switched between servers on the fly. Second reason would be to reduce the number of spare drives needed, eg you don't need 2 parity drives per server, but more like 2 for every 10 drives.

 

Data vs RAM vs CPU vs Transfer Speeds, do you need a lot of storage, but also a lot of ram or a lot of cpu? you want network storage + many physical servers. Is the CPU/RAM needs limited to a small subset of your total data, eg database vs mass storage if videos? Maybe local storage is better.

 

Expanding storage is easy enough, if you just have a few towers, fill them with 8TB HDDs, and SSDs fit almost anywhere.

 

What model of IBM SAN costs £7k?

Posted
Guys I'm starting to look at upgrading some of our servers next financial year and I'm considering the possibility of a single disk server with no RAID configuration.As budgets get tighter and tighter I can probably afford to put in place two single disk Hyper-v host servers

 

...

 

We actually have 3 SANs.

 

Well that escalated quickly.

  • Thanks 4

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...