We’ve just added a new Hints section to our website. To start with, check out our mini-articles on hard drives, FaceBook, DNS, Dropbox and Laptops. We’ll be adding lots more over time, which we will announce through this blog. We’re intentionally keeping the hints separate as they serve different audiences – this blog should be readable by a non-technical person, whereas the hints are designed for a semi-technical or technically-curious person. The hints will also be much shorter and with less justification and consideration of options. Anyway, if there’s a hint or topic you’d like us to cover, please let us know.
We work on a lot of IT projects, and we especially enjoy working on integration projects through our FlowLabs unit. Integration, workflow and custom programming work nearly always have compelling, measurable ROI for clients. Most clients have a strategy for getting the IT project work they need within their budget. We’ve seen quite a few different strategies—some work for the client, and some don’t.
If you want to minimise the costs for your next IT project, try thinking about how to minimise the project risks. Beyond a certain level of complexity, there are risks for both the client and the IT services provider. The more unknowns there are, the higher the level of risk. The good news is risk doesn’t have to mean that the amount of money you pay goes through the roof. We recommend three rules for getting quality IT project outcomes from your consultants:
- Disaggregate (break up) projects as much as possible. Each time you make the target smaller, you decrease your risks. Typically, the more often you do this, the less risky subsequent parts of the project become.
- Don’t assume that a fixed price is the best way to avoid being ‘ripped off’. We never offer fixed price where it’s not in our client’s interests, but it’s quite common in the industry to heavily load fixed prices to militate against risk. A good IT services consultant will help you manage the risks and be able to suggest the pricing structure that will be most advantageous to you.
- Work with an agile process. Agile processes emphasise discovery, lightweight solutions and iteration over heavy planning and over-managing. Applying agile methodologies to a project gets you results faster, which again decreases risks and costs.
- Think long term. You might have a need now, but what are your needs going to be in 5 years time? You don’t have to code that now, but knowing where you’re heading makes an enormous difference to the design decisions we collectively make now.
- The single best thing you can do to reduce risks is reduce complexity. The single best thing you can do to reduce complexity is aim towards one database. You don’t have to get there all at once, but if you’re heading in that direction, things will get easier for you over time rather than harder. Systems that look like they can talk to each other now often don’t work properly, or can’t when you swap programs or are so convoluted your staff hate them. If you’re looking for a real competitive edge and one that will last for a long time, this is a great place to start.
Most businesses want two backups for their data: an onsite system and an offsite system. (It’s worth pointing out that this point that RAID is not backup!) One convenient way of doing offsite backups is over the internet. Data can be backed up from one branch office to another, or into a data center, or your IT services provider might provide a backup server at their premises. We’ve set up all three of these types of offsite backup for clients multiple times over the years.
Off-site backup can only works in practice if you have a differential backup system. This means instead of a whole copy of your data set being copied remotely, only the difference between today and yesterday is. This is usually, but not always (see below), a small fraction of your data set.
On average, a corporate ADSL connection could theoretically cope with about 300GB a month of off-site backup. Theoretically. The obvious problems are surmountable – that your ISP doesn’t give you your full bandwidth (change ISPs) or that your quota isn’t high enough (increase it). The less obvious problem is that relying on differentials is only something that works most of the time, not all of it. So what happens when you upload a DVD to your server, or you sort a whole bunch of files into different folders, or you simply have lots of data changing regularly? Your off-site backup system starts to play catch-up. It could be months before it gets back keeping your stuff in sync. It’s trivially easy to get to the point where the estimated time of being up to speed is ‘never’. And who is monitoring all this whilst it’s happening?
Note that even backing up essential files on a single computer daily would overpower any (non-NBN) consumer-grade internet connection in this country if you didn’t use differentials. Differential backup is more complicated to set up (CoreServer does it out of the box), so off-site backup is either skipped (which might not be within your acceptable risk tolerances) or done poorly with a system that isn’t designed properly, tested or monitored. In short, off-site backup is technically complex and requires careful planning and a frank discussion about risks.
As a result of all these limits and complexity, it’s not actually possible to say you’ll be OK with only “x” GB of data. In fact, the only number you can truly be safe with is an amount you know you can back up, in full, every day if you had to. On a typical ADSL connection with a more expensive plan, this means between 3GB and 20GB of total data storage, the latter only if you have a fantastic internet connection right next to a telephone exchange with ADSL2+. As you can see, online backup is no panacea.
There are many online backup solutions out there and you may be tempted to try them. But most of them are designed for radically different markets – e.g. much of the USA, Japan or south-east Asia where bandwidth is much greater. In Australia, the limit is related to ADSL in Australia and its costs. Cable is no better (for similar, but different reasons). You may have noticed that upload is much slower than download for ADSL connections. That’s what the “A” in “ADSL” stands for – asynchronous – not the same speed. Standard ADSL uploads top out at about 1MBit/sec, which equates roughly to 300GB a month, assuming you get the full speed all the time, and your backup demands are smooth (which they won’t be). ADSL2+ tops out at around 2Mbit/sec. We see many clients with ADSL upload speeds well under 512kb/sec. You can check this yourself at SpeedTest to see how your connection fares (make sure you shut off anything else using the internet at the time to get a fair reading). Once you move beyond standard ADSL connections, prices for high-upload services move quickly toward the stratosphere. (A comparison from Whirlpool is here.). This is true corporate internet. We can help you attain this, but most of our smaller and some of our larger clients are shocked at the cost, particularly if they’re just looking for backup. This is where we’ve implemented a couple of cheeky solutions for people in the past:
- Using point-to-point wireless between nearby offices to provide most (but not all) of the benefits of off-site backup cheaply
- Swapping off-site backup storage (with encryption) with nearby offices
- Doing manual once-off data “seeding” then relying on online backup for reliably small differentials
If you’re not willing to spend huge amounts of money on bandwidth, but still need offsite backup, the most cost-effective solutions are disk-based backups or a semi-automated hybrid of internet and disk-based backups. As the NBN expands, this will also bring down the cost of doing large backups over the internet. We have a bunch of solutions up our sleeve here that are not quite ready for prime time, but stay tuned or contact us to find out more.
N.B. This post did not explore non-online forms off off-site backup. In short, these are not a panacea and require considerably more manual effort to verify they actually work. That, however, is the subject of another post.
Servers are part of the core business toolkit of nearly all small businesses these days (and all larger ones). Most businesses intuitively know when they need one, and know what they need it for. The server is like a hub, helping them share files, do backup, manage their internet connection, and provide email. The good news is that it’s pretty hard to buy yourself a server that doesn’t offer these basic functionalities.
Like any piece of technical equipment, it’s the specifications that can be confusing. Do you really need that extra hard disk, power supply or CPU? If you’ve got an IT provider you can trust, they should be able to help you navigate which options you need for your business. Here’s three things you should think about:
- Think about growth. We design our servers to be replaced after three years, which is good practice to lower your malfunction risks. How many staff will you have in three years time? How many files will you need to store? How many desks will you need to serve? What volume of email? What volume of network/internet traffic?
- Think about backup & recovery. There’s backup and then there’s backup. Unfortunately, not all backups are created equal. More worryingly, a lot of backup systems simply don’t work. At the basic level, a good backup solution will have RAID and be monitored. Your IT provider should be able to discuss the pros and cons of the various types of RAID & storage systems available (hardware vs software, magnetic vs SSD, write-safe vs non-write-safe caches, battery backup, etc). You should also be sure that your backup solution scales with your growing requirements. For example, an offsite backup solution over the internet isn’t going to work if your daily changes exceed your daily upload bandwidth (which is very easy to do in Australia.)
- Think about maintenance & monitoring. IT departments in large companies help keep servers in tip-top operating condition, doing scheduled maintenance, monitoring the health of key hardware, checking systems aren’t compromised, and other vital caretaking work. This proactive approach ensures maximum uptime and minimum risk. In the small business environment, unmaintained servers pose a real threat to data and productivity. We see small businesses out there all the time with ticking time bombs of data unreliability. This maintenance and monitoring should be very detailed when it comes to your storage stack – physical drives, controllers, RAID volumes, logical volumes and backup. This is detail-oriented work and you need someone who sweats these details looking after your precious data. Your IT provider should offer you a maintenance and monitoring program to ensure your most vital piece of hardware is performing at its best. We think this is so important that we never sell a CoreServer without the peace of mind of a management program.
On top of all that, you should ask what’s not included with your server quote. Do you need a separate router? A separate VPN server? A separate terminal server? Are these things properly specified or afterthoughts? (e.g. We see many great servers sitting behind totally inadequate consumer-grade ADSL router/modem/wifi devices). Separate licenses? Is installation included – if so, to what extent? Will you get training? Or will the provider move on to their next client after they’ve finished with you? There are no right answers to these – it all depends on your business objectives and desired commercial relationship. But you will almost certainly be up for a shock in the future if you or your provider don’t raise these things upfront.