- the ability of something to be changed in size or scale
- the ability of something to be used or produced in a range a capabilities
Tech people love to talk about scalability. They especially like to use the term with school administrators when talking about how much money the IT department will need in future years to enable them to properly scale the technology infrastructure required to support the academic and operational goals of the school. IT directors will often ask themselves and vendors if “such and such a system” can scale with future needs. Typical of such systems are network performance, disk storage space, speed of applications, processor power, and other QOS (Quality of Service) standards.
An unspoken assumption regarding scalability is that it almost always means scaling up: more storage, faster CPUs, more services, more users, more devices ad nauseam. But in truth scalability includes the ability to contract as well as expand. Some technology scales very easily in both directions while others, as we will see, do not.
Here are some factors to consider in the scaling technology discussion.
- Size Matters. Systems are scalable when they work for all sizes of people.
- Technologists sometimes forget that children have smaller hands and legs than adults. Small children are often asked to adapt to adult sized computer keyboards, pointing devices, and furnishings. Without fully developed their fine motor skills, and keyboarding and pointer skills may be developmentally difficult for them.
- This is why computer tablets are especially useful with younger learners where the direct, tactile manipulation of screen objects is both more intuitive and developmentally appropriate.
- Adults, on the other hand, may find that smaller devices such as smartphones are more cumbersome and difficult to read. Perhaps this is in part why smartphones are trending towards larger screens.
- The Cost of Unused Capacity. Systems are scalable when they are in the “goldilocks zone” not too big, not too small.
- In many areas of technology cost per unit of measure has fallen over time.
- For example, the cost of additional random access memory (RAM) has fallen from $189/G in 2005, to $5.5/G in 2013; the cost per gigabyte of hard disk capacity has fallen from $1.24/G in 2003 to $.05/G in 2013.
- These falling prices have been a boon for technology department budgets, even counting the unintended consequences of bloated software and storage of unnecessary and redundant files. [See Delete This Post.]
- But just because RAM and storage is cheap doesn’t mean schools should not exercise due diligence in stewarding these resources.
- On the other hand, Internet bandwidth in the United States is quite expensive compared to other countries (see table from newamerica.net.
- For example, in Seoul a gigabyte internet connection will set you back only $31.47 per month, compared to almost $1000 per month for the same level of service in Lafayette, LA. Your local service likely likes between these extremes, and competition is gradually increasing in American cities.
- Be wary of long-term contracts that may lock you in to high prices for unused bandwidth.
- Some ISPs offer so-called “burstable” bandwidth which purport to allow customers to get faster internet speeds on an as-needed basis. But the fine print in these contracts often reveals that they are not what they are advertised to be. (see links at end of post)
- In many areas of technology cost per unit of measure has fallen over time.
- Standardization vs. Personalization. Systems are scalable when personalization is the responsibility of end users, not the IT department.
- The BYOD model of provisioning student computers coupled with the influx of consumer devices (smartphones, e-readers, and so on) means that tech departments now face a much wider range of support issues than when everyone used the same computer.
- Support of BYOD models scale best when management of the physical devices brought to campus is limited to cloud-based services and Internet browsers with zero or very few locally installed applications.
- When it comes to computers that are purchased by schools for use by employees, IT departments must weigh the support costs of standardization versus the desire for personal choice and configuration of devices from faculty and staff accustomed to this in their personal technology use.
- Personalization does not scale well if it the model of your help desk is “the do it for me” rather than “teach me how do it for myself” department.
- Finding the right tension between personalization and standardization is part of IT’s mission.
- Sometimes, just say “no.” Systems are scalable when they do not overwhelm human capacity.
- Human capacity for change is less than the breadth of opportunity in educational technology. I have witnesses the stress in many of the teachers with whom I consult, teachers who are dealing with (a) the introduction of a 1-1 device initiative, (b) the adoption of Common Core State Standards), (c) pressure from lawmakers and regulators as manifested in high-stakes testing and anti-union rhetoric, (d) increasingly litigious parents, (e) decreases in school funding, and (f) increases in school enrollment.
- Businesses who grow to fast fail. (Remember the original People’s Express Airline debacle.) Teachers who are asked to do too much too fast burn out.
- Uh, that was unexpected. Systems are scalable when their core functionality remains strong over time.
- A great decision made in 2009 may turn out to be a lousy decision in 2014. A great 1.0 version of a product may, by the time it reaches version 5.0 or 6.0 be so unrecognizable, bloated, and feature laden that simplicity, ease of use, organizational utility are missing and the cost/benefit ratio is negative.
- One could argue that long in the tooth applications such as MS Office have reached this stage and that cheaper, basic word processing, spreadsheet, and presentation application make more sense for schools. But core applications such as these are “sticky” and difficult to dislodge from user allegiance and institutional inertia.
- When considering scalability it is important to determine what are the critical features in a product or service that you wish to see develop capacity over time and which are not.
- “Feature creep” is a well-known phenomenon in software development and sometimes it is the end users who are victimized by it.
- Software companies and service providers have profit incentives to continually innovate and generate additional revenues by convincing you that current technologies are inadequate. Indeed, there often seems to be an unholy alliance between software and hardware companies to use consume more and more CPU cycles and hard drive space rather than optimize software to be more efficient.
- You can check out any time you want, but you can never leave. Systems are scalable when you can easily get stuff back that you put in.
- Technology partners are like banks, banks in which you deposit assets (data), the privilege for which you are charged a fee.
- Should the bank go out of business or another one offer you a better deal, you will want to easily get your data back. This is often easier said than done.
- Assume that you contract with a company to host your school’s student information system (SIS). The company stores your data in a standard database format such as SQL. So far, so good. Over time you ask for modifications that required modifications to the database as well as special report formats which the hosting company creates using non-standard, proprietary software. Ten years later you decide you want to switch to another SIS only to find that the customizations and reports will make it difficult to get your data out and proprietary report formats and coding are not owned by you. Do you stay with the old system despite its flaws or move to the new system?
- Economists might characterize this problem as a sunk costs problem, in which the investments you have made in the first system are not recoverable and you need to take the hit and move on. But many people do not think like economists and continue to have faith in and use systems that don’t meet their needs simply because they have put so much into it.
- This technology is rated PG-13. Systems are scalable when they work across different age groups, cultures, and skill levels.
- One might think that building user interfaces that are “intuitive” would be, well, intuitive. But this is not the case. Intuition is highly dependent upon age, experience, culture, and patterns of problem solving and thinking. When first introduced, the computer mouse was not intuitive to many users. They would pick it up, move it through the air, run it across the computer screen, or make other errors that we perceive as humorous now.
- If you must make significant changes to computing systems that are not intuitive for 98% of the users of the hardware or software, I suggest you look rethink your rollout plan.
- A Greener Shade of (e)mail. Systems are scalable when their power requirements stay the same or decrease even as more users or functionality is added.
- Energy efficiency of computing devices roughly doubles every year and a half, a trend that needs to be factored when considering a school’s investment in hardware, especially when such devices are part of an always on, always available Internet. A single device may not consume that much power but in aggregate all of the devices in your school do.
- When calculating your school’s carbon footprint, you must include a portion of the power consumed by associated hosting services, and even estimates of recharging of school devices that may happen at home.
- Gilding the lily. Systems are scalable when perfection does not become the enemy of “good enough.”
- Google has reported that it wants to get 90% of all MS Office users for whom Google Drive will do what they need, leaving the last 10% who need a desktop application to continue to use Office. It may well be that their analysis of the functions available in Google Drive are in fact sufficient for 90% of users, but I suspect they underestimate the amount of effort it will take to change people’s habits.
- It is the responsibility of the tech department to spend money wisely. People making an in-house newsletter do not need Adobe InDesign to accomplish that. And most users do not need Microsoft Office to write memos or letters, make simple worksheets, or adequate presentations.
Have your own ideas about scalability? Leave a comment!
Resources on burstable bandwidth: