If you’ve sat through a training seminar recently, you’ve probably seen a slide talking about software-defined anything or software-defined everything. Or you may have seen the acronym SDx and wondered what it is and where it’s come from. So let’s have a look at what they’re talking about.
Basically, what we’re looking at is using software to control different kinds of hardware, and then to make that software able to control multiple-component hardware systems. With the growth of the Internet of Things (IoT), it makes sense to start thinking about being able to create rules that are implemented in software that can be used to control a myriad of different types of devices.
At the moment there are a number areas using software-defined technology. For example there’s Software-Defined Storage (SDS), which seems to apply to all sorts of storage software, particularly virtualization software. Different vendors use the term loosely for different things. Software-Defined Networking (SDN) is where network devices are programmable and so networks themselves are more dynamic. Again, it’s a term that’s used by different vendors for different things.
Software-Defined Storage Networks (SDSN) is an attempt to virtualize storage networks by separating the actual physical network from its controlling software. A Software-Defined Hypervisor (SDH) seems to refer to virtualizing the hypervisor layer and separating it from its management console. And finally, there’s Software-Defined Infrastructure (SDI) aka Software-Defined Data Centre (SDDC), which is an aspirational concept where data centre services are controlled by policy-driven software.
Two things probably leap to mind about now. Firstly, this seems a lot like marketecture! We’ve seen this before, where vendors are really selling us an idea of something rather than it being a tangible reality. We are very much in the early days of this sort of thing. The second thing is that this is not directly linked to mainframes. This is VMware’s ideas – as well as a huge number of other companies.
Having said that, of course, the newer hybrid mainframes from IBM will be able to make use of this technology as it becomes available in reality. Also, Gartner reckons that SDx is one of the major disruptive technologies to watch. It makes it easier to scale up existing architecture and even try out different architectures. Also it makes it possible to tune networks, matching network performance to workloads. And, of course the main selling points are flexibility, agility, security, and price.
IBM’s Smarter Computing blog has an interesting blog by Shamin Hossain called “Software defined everything: When a data center becomes soft”, which can be found at http://www.smartercomputingblog.com/software-defined-environment-2/software-defined-everything/.
Clearly the prefix ‘software-defined’ is one that we’re going to hear a lot more about this year.
Showing posts with label storage. Show all posts
Showing posts with label storage. Show all posts
Sunday, 23 March 2014
Sunday, 9 September 2012
Back-ups, archives, and Pogoplug
Storage is always an issue. Whether you’ve got petabytes of it on your mainframe or just 16GB on your smartphone or tablet, the issues are always the same. You want speedy access to the data you’re using, and you want somewhere to store the data you’re not using at the moment – but might want to use it later.
Speeding access to data can be helped in a number of ways – storing the data contiguously, using indexes to speed up access to individual fields, using shorter paths from the storage medium to the processor, keeping recently used data in cache storage in the likelihood it will be used again. Human ingenuity has come up with lots of workarounds to ensure that you get the data as quickly as possible.
But there’s more of an issue with archived data. I could have immediate access to it, but then it would cost as much to store as live data. I could store it on some cheap medium and find it takes a week to get it back to a state that I can use. And something along those lines is used by many archiving organizations. Those tapes are stored inside a mountain somewhere, and the data can be restored – but not quickly.
Depending on the kind of business you’re in, someone has made a decision about costs and storage retrieval speed. If your company is in banking or insurance, that data might get back much faster than if your organization doesn’t promise its customers fast access to old data.
The advent of cloud computing added another layer of complexity to storage. You could store your data on the cloud – which, in psychology terms, is called a nominalization. You call it something, which seems to have meaning, but isn’t a concrete real thing – you can’t put it in a wheel barrow. You hear politicians use nominalizations all the time. I’ll vote for prosperity and world peace – but what does that actually mean? Prosperity for someone from a third world country might seem quite close to poverty to me! And world peace can always be achieved by having a tank on everyone’s lawn. Your idea of what is meant by a nominalization may be completely different from the idea held by the next person!
Anyway, I digress. So nowadays, you can store your data on the cloud. You don’t know where it’s physically stored you just know how to access it. And I assume much the same models are used for the price of cloud storage – fast access is dearer than slow access. Amazon came up with, what they call, Amazon Glacier. This is a cheap service for archived or backed up data that you don’t access very often. In line with that pricing model, retrieval can take several hours. Amazon Glacier only charges organizations for the storage space they use.
I’ve been using a Pogoplug for over two years and I’ve mentioned it in this blog a few times. A Pogoplug started life as a small device that plugged into your router and allowed you to plug in memory sticks that you could access from a browser anywhere. The company has recently expanded its cloud offerings and has done a deal with Amazon to offer cloud storage at a very competitive rate. The solution isn’t for mainframers, but makes sense for individuals and small companies.
The Pogoplug Web site at http://pogoplug.com gives all the price plans they’re offering. So, a small organization would be able to back up files from their computers, smartphones, and tablets to one secure location. All the company has to do is buy a Pogoplug device or download the Pogoplug Server software and run it on a Windows or Mac. They can then back-up and archive their files – and we’re talking terabytes of space. Users can continuously copy all or part of their data to Pogoplug’s secure offsite cloud. Any changes they make to their original folder will be synchronized automatically.
It seems like a useful addition to the all-important back-up and restore choices a company has to make.
Speeding access to data can be helped in a number of ways – storing the data contiguously, using indexes to speed up access to individual fields, using shorter paths from the storage medium to the processor, keeping recently used data in cache storage in the likelihood it will be used again. Human ingenuity has come up with lots of workarounds to ensure that you get the data as quickly as possible.
But there’s more of an issue with archived data. I could have immediate access to it, but then it would cost as much to store as live data. I could store it on some cheap medium and find it takes a week to get it back to a state that I can use. And something along those lines is used by many archiving organizations. Those tapes are stored inside a mountain somewhere, and the data can be restored – but not quickly.
Depending on the kind of business you’re in, someone has made a decision about costs and storage retrieval speed. If your company is in banking or insurance, that data might get back much faster than if your organization doesn’t promise its customers fast access to old data.
The advent of cloud computing added another layer of complexity to storage. You could store your data on the cloud – which, in psychology terms, is called a nominalization. You call it something, which seems to have meaning, but isn’t a concrete real thing – you can’t put it in a wheel barrow. You hear politicians use nominalizations all the time. I’ll vote for prosperity and world peace – but what does that actually mean? Prosperity for someone from a third world country might seem quite close to poverty to me! And world peace can always be achieved by having a tank on everyone’s lawn. Your idea of what is meant by a nominalization may be completely different from the idea held by the next person!
Anyway, I digress. So nowadays, you can store your data on the cloud. You don’t know where it’s physically stored you just know how to access it. And I assume much the same models are used for the price of cloud storage – fast access is dearer than slow access. Amazon came up with, what they call, Amazon Glacier. This is a cheap service for archived or backed up data that you don’t access very often. In line with that pricing model, retrieval can take several hours. Amazon Glacier only charges organizations for the storage space they use.
I’ve been using a Pogoplug for over two years and I’ve mentioned it in this blog a few times. A Pogoplug started life as a small device that plugged into your router and allowed you to plug in memory sticks that you could access from a browser anywhere. The company has recently expanded its cloud offerings and has done a deal with Amazon to offer cloud storage at a very competitive rate. The solution isn’t for mainframers, but makes sense for individuals and small companies.
The Pogoplug Web site at http://pogoplug.com gives all the price plans they’re offering. So, a small organization would be able to back up files from their computers, smartphones, and tablets to one secure location. All the company has to do is buy a Pogoplug device or download the Pogoplug Server software and run it on a Windows or Mac. They can then back-up and archive their files – and we’re talking terabytes of space. Users can continuously copy all or part of their data to Pogoplug’s secure offsite cloud. Any changes they make to their original folder will be synchronized automatically.
It seems like a useful addition to the all-important back-up and restore choices a company has to make.
Subscribe to:
Posts (Atom)