Amazon’s Elastic Load Balancer is a Strangler

The Strangler is an extremely effective technique for phasing out legacy systems over time.  Instead of spending months getting a new system up to parity with the current system so that clients can use it, you place The Strangler between the original system and the clients.  

The Strangler passes any request that the new system can’t handle on to the legacy system.  Over time, the new system handles more and more, the legacy system does less and less. Most importantly, your clients won’t have to do any migration work and won’t notice a thing as the legacy system fades away.

A common objection to setting up a Strangler is that it Yet Another Thing that your overloaded team needs to build.  Write a request proxy on top of rewriting the original requests! Who has time? 

Except, AWS customers already have a fully featured Strangler on their shelf.  The Elastic Load Balancer (ELB) is a tool that takes incoming requests and forwards them on to another server.

The only requirement is that your clients access your application through a DNS name.

With an afternoon’s worth of effort you can set up a Strangler for your legacy application.

You no longer need to get the new system up to feature parity for clients to start using it!  Instead, new features get routed to the new server, while old ones stay with the legacy system.  When you do have time or a business reason to replace an existing feature the release is nothing more than a config change.

Getting a new system up to parity with the legacy system is a long process with little business value.  The Strangler lets your new system leverage the legacy system, and you don’t even have to let your clients know about the migration.  The Strangler is your Best Alternative to a Total Rewrite!

Probing Questions For Executives Approving A Rewrite

A while back, I wrote a guide for developers preparing for a system rewrite.  This time I’m talking to the other side of the table; executives being asked to approve a system rewrite.

Here are three probing questions to help you and your developers dig into the realities of a rewrite and gain confidence in the project:

  1. What are the client’s pains, and how does the rewrite solve them?
  2. How will you support current users during the rewrite?
  3. What lessons have you learned from the original system, and how will you prevent repeating the same mistakes?

What are the client’s pains, and how does the rewrite solve them?

There is no business value in a the next generation of your software.  Have your developers put in the work to understand how the current system causes client pain?  

Do they know the frustrating pain of needing features that can’t be delivered? The maddening pain of redoing work destroyed by a bug? The teeth grinding pain of slow systems that steal minutes with deadlines hours away?

Which pains are they addressing, and how much of the system do they need to rewrite to give clients some relief?

Pain provides focus and a metric for success. Without a specific pain goal, the project will metastasize and grow into “replace everything”. 

How will you support current users during the rewrite?

Feature freeze is the most common answer, but it should not be acceptable.  Neither is only patching critical bugs.

One solution that has worked well for me: The Senior proposing the rewrite gets to architect and oversee the new solution, but all the coding will be done by the juniors.  The senior will be responsible for fixing bugs and implementing new features on the legacy system.  

With that much skin in the game, and no juniors doing junior things, you’ll be amazed at how quickly the legacy system stabilizes. I have seen this be so successful that the rewrite gets canceled, and the work written off as a moral boosting learning experience.

What lessons have you learned from the original system, and how will you prevent repeating the same mistakes?

This is really a question of ownership.  Are your developers ready to acknowledge and own their mistakes?

Are their expectations for the future realistic, or do they expect the shiny new technology to prevent them from making the same mistakes?

Going from monoliths to micro services, micro services to serverless, or serverless back to a monolith won’t help if your developers are repeatedly implementing a fundamentally flawed design.

Rewrites can destroy your company

Companies have been destroyed by endlessly stalled rewrites.

Work with your developers to answer these questions.  It’s likely that you’ll find a much less risky path than the full rewrite.  If you do agree to the rewrite, asking questions will give you confidence that your developers have a realistic plan that they can deliver.

Questions To Ask Before Automating

The prospect of automating manual tasks emits a siren song to most developers.  Like a siren, the call often leads you straight into disaster. Best intentions often end up leaving companies with code that’s more expensive to maintain and less useful than human labor.  Reports and tasks become a leaky faucet for productivity.

Here are six questions to ask yourself, or a developer, before dancing to the automation music:

How often is the task likely to change?

Weekly Business Intelligence reports change monthly, monthly ones change every quarter, and quarterly ones change every year.  They are never stable enough to be worth automating by an outside developer. This is why BI tools that let non-technical users semi-automate reports are a 5 billion dollar industry.

On the other hand, regulatory and compliance reports are likely to be stable for years and make great targets.

If a task won’t be executed at least 10 times between changes, it probably won’t be worth automating.

How long is the task likely to continue?

Some tasks are likely to continue “forever”.  Archiving old client data, scrubbing databases and other client onboarding/offloading tasks fall into this category.

Some tasks are never going to come up again.

If a task won’t be executed at least 5 more times, it probably won’t be worth automating.

How much human effort does the task consume, and for whom?

You can automate turning on the office lights in the morning with a motion detector, but it won’t pay off in terms of time saved from flipping a switch.

How much of an interruption is doing the task?  Turning on the lights on your way in the door isn’t an interruption for anyone.  Phone support manually resetting a user password isn’t an interruption, but having the CFO process refunds for clients is a giant interruption.

Even if the reset and refund are both a single button click that takes 15 seconds, pulling the CFO away is a much bigger deal.  Also the context switch for the CFO will be measured in minutes because she’s not processing refunds all day long.

Use a sliding scale based on time and title.  For entry level, don’t automate until the task takes more than an hour per person per day.  For the C-Suite anything over 5 minutes per day is a good target.

How much lag can automation save your clients?

Clients don’t care how long the task takes, they care about the lag between asking and receiving.  It doesn’t matter that processing a refund only take 5 minutes if your team only processes refunds once a week.

If the client lag is more than a day, consider automating.

Is the Task a real process, or are you cleaning up the effects of a bug?

Software bugs can do all sorts of terrible things to your data and process, but after the first couple of times, the damage becomes predictable and you’ll get better at fixing the damage.  

Automating the fix is one way of fixing the bug.  That’s how bugs become features.

If you don’t want to make the bug an official part of your software, don’t automate the fix.

How common and expensive are mistakes?

Mistakes are inevitable when humans are manually performing routine tasks.  

Mistakes are inevitable when developers first automate a routine task.  Assume that developer mistakes will equal one instance of a manual mistake.  

For an automation to save money you have to expect to prevent at least 2 manual errors.

As an equation:

[Cost to Automate] + [Cost of a mistake] < [Cost of a mistake] * [Frequency of mistakes]

Because the cost of mistakes is relatively easy to quantify, tasks with expensive mistakes are usually automated early on.

Conclusion

Developers always want to automate things, sometimes it pays off, sometimes it’s a mistake.

If you ask these six questions before automating you’re much more likely to make the right choice:

  1. How often is the task likely to change?
  2. How long is the task likely to continue?
  3. How much human effort does the task consume, and for whom?
  4. How much lag can automation save your clients?
  5. Is the Task a real process, or are you cleaning up the effects of a bug?
  6. How common and expensive are mistakes?

When Automation Wastes Time and is Expensive

By the time a company gets to midsize, there will be a pool of people to perform ad hoc, automatable tasks.  Most of these tasks involve manually generating reports for someone else. To someone who writes software, this can seem infuriatingly inefficient and wasteful.  A sign of tech debt and bad software! 

Counterintuitively, it is extremely difficult to automate reports and actually save the company time or money.

Imagine Alice, the CEO, has Bob, in accounting, make a weekly report by hand pulling numbers from 3 different systems and plugging the data into a spreadsheet.  Alice doesn’t care how Bob generates the report, so Carol, a developer, decides to help Bob by automating the report.

To keep the numbers simple, Bob earns 75k/year (~$36/hr) and Carol earns 150k/yr (~$72/hr).  Let’s say it takes Bob 1 hour/week to generate the report, making it cost $1,872/year.  

All the data is in databases, and Bob is doing specific steps in a specific order.  He’s miserable because making the report is dull, error prone, repetitive and endless.  Worst of all, the report is meaningless to Bob! He’s generating it for Alice, because he’s capable, understands what the numbers mean, and has access.  Carol is Bob’s friend and wants to help!

Carol’s time is more valuable than Bob’s, so she can only spend 3 days automating the report for the company to break even.  That includes testing, formatting, and deployment.

Let’s assume Carol is able to automate the report on time.  The report shows up in Alice’s inbox every week, same as before, but Bob is able to work on other things.  Break out the champagne!

Except, Alice needs changes to the report twice a year.  When Bob was generating the report she could just tell him and it would take maybe 30 minutes to change the process because Bob works in accounting and understands all of the numbers.  There was no lag between the ask and the update, Bob would make the updated report next time.

Now that Carol has automated the report, Alice will have to talk to Bob, and Bob will have to talk to Carol.  Alice can’t ask Carol directly because Carol doesn’t work in accounting and doesn’t know what the numbers mean.  Additionally, Carol doesn’t spend her days working on reports so she’ll have to get up to speed, talk to Bob and make the change.  Let’s say it takes one full 8hr work day.  

As the CEO, Alice’s changes are important, but probably not important enough to pull Carol out of her current work.  It usually takes 2 report cycles to get to the top of her list.

Now the automated report costs $1,152/yr to maintain, and it’s not the exact report that the CEO wants 4 weeks, or 7% of the time.  4 out of date reports easily outweighs the $720 in savings and Alice will probably have Bob manually generate the report when the code is out of date.  Finally, the sunk cost of building the software will result in Carol having to maintain the software going forward, even though the software generates no real value.

The conflict between “can be automated” and “can’t be done profitably” often results in DBAs become a report generation team.  It’s also the impetus behind business intelligence tools; find a way to get developer style automation, without using a developer.

I’ll go over a developer’s “should I automate this task” framework next time.

A Tale of Two Failures

It was the best of times, it was the worst of times.  The company hired contractors to do maintenance work so that the full time developers can write a new system, the company hired new developers to write a new system stuck original developers with the maintenance work.

When software has been bankrupted by tech debt, a common strategy is to start over.  Just like financial bankruptcy, the declaration is a way to buy time while continuing to run the business.  The software can’t be turned off, someone has to keep the old system running while the new system is built. Depending on your faith and trust in your current employees you will be tempted to either hire “short term” contractors to keep things running, or hire a new team to write the new system.

The two tactics play out differently, but the end result is the same, the new system will fail.

Why Maintenance Contractors Fail

Contractors fail because they won’t have the historical context for the code.  They may know how it works today, but not how it got there. Without deep business knowledge their actions are limited to superficial bug fixes and tactical features.  The job is to keep the lights on, not to push back on requests or ask Who, What, Where, When, Why.

Since the contractors are keeping the lights on for you while current employees build a new system, they aren’t going to care about the long term quality of the system.  It’s already bad, it’s already going away, why spend the extra time refactoring.

Contractors won’t be able to stabilize the system and buy your employees enough time to build the new system.  Equally bad, sometimes the contractors *will* stabilize the old system which becomes an excuse to cancel the new system and not restructure your debts.

Why Hiring a New Team Fails Too

The flip side is to leave the current team in place and hire a new team to write the new system.  This is super attractive when you suspect that the original team will recreate the same mistakes that led to disaster.

This tactic fails because the new team won’t have the historical context for the code.  Documentation and details will be superficial and lack all of the critical edge cases. All of which delays the project.  Meanwhile, your original team will become extremely demoralized.  

If they couldn’t keep things running before, wait until you tell them their whole job is to keep things running.  The most capable members will quickly leave and you’ll have to pull developers from the new team onto the old system.

You’ll end up pulling the new developers onto the old team and praying they don’t see it as bait-and-switch.

Don’t split on Old vs New

Both tactics fail because the teams are divided between old vs new.  In the end it doesn’t matter who is on which team, because you need old *and* new to be successful.

Keeping the team together ensures that everyone has context into how the system works and has skin in the game.

Instead of splitting old vs new, find a way to split the system in half.  Have each team own half the old system responsibilities, and one of the two new systems.  This gives everyone a stake in keeping the old system alive, a chance to work on the new system, and reduces the business risk of either team failing.

Once you find a way to split the responsibilities in half, you might even find a way to make iterative improvements, it’s your Best Alternative to a Total Rewrite!

A Guide to Prepare for a System Rewrite

I often talk about the Best Alternative To A Total Rewrite.  The idea that you should know the alternatives to a rewrite before making a decision.

Even when there is no alternative to a rewrite you still need to work through the common problems that cause the project to fail.

These open ended questions are designed to guide your preparations:

How will a rewrite solve your users’ pain?

Where is the current system failing your users?  How will a rewrite fix those problems? You and your fellow developers may hate the current codebase, but that isn’t a compelling argument.

Clients don’t want the Next Generation of your system.

Are there really no incremental improvements?

Now that you know what user pain you are trying to solve, is there really no way to resolve the problem in the current codebase?  

If the system is slow, is there really nothing you can do to make it faster?  Maybe a rewrite would make things so fast that users won’t notice any lag. But if a refactor would reduce frustrating lag by 50% in a month, you should refactor.

Once you’ve focused on your client’s problems, reducing pain today is preferable to fixing the problem 6 months from now.

Do you have to recreate the existing system before adding new features?

After bugs and latency, the number one issue with old codebases is that it’s hard to add new features.  The original code was designed to expand in one way, and now it needs to expand differently. A rewrite will let you design a system to meet those new requirements.

What about a new system that is nothing but new features?  Can you find a way to make a new system that is a client of the old system?

A rewrite requires you have to recreate all the features of the existing system before adding the new features that create new business value.

How will you ensure that the new system doesn’t have the same problems as the original system?

The forces that caused the original system to go off the rails are going to work against the rewrite too.  

Management didn’t want to spend time on unit tests?  They still won’t.  

Constantly changing business requirements?  Just wait until the demands aren’t constrained by the existing code.

Inadequate resources, logging, metrics and other tools?  A rewrite won’t make any of that better.

After a brief honeymoon a rewrite will face more pressure than the original system.  You need a plan to keep all of those external forces at bay.

What’s the release plan?

Replacement is not a release plan.  How is the rewrite getting into production?  Can you find a way to run the two systems in parallel?  Can you replace things a piece at a time?

You will have to release the new system eventually. “When it’s done” takes forever.

Conclusion

When you propose a rewrite, you need to be able to answer these questions to your teammates and managers.  Bringing the questions and answers up in your initial pitch will show that you’ve thought things through.

Don’t proceed until you can answer all five questions!  You just might happen to discover an alternative to a total rewrite.  Even if you don’t use it, you can move more confidently once you can speak to the alternatives.

Dead Database Pixel Tracking

Pixel Tracking is a common Marketing SaaS activity used to track page loads.  Today I am going to try and tie several earlier posts together and show how to evolve a frustrating Pixel Tracking architecture into one that can survive database outages.

Pixel Tracking events are synchronously written to the database.  A job processor uses the database as a queue to find updates, and farms out processing tasks.

Designed to Punish Users

This design is governed by database performance.  As the load ramps up, users are going to notice lagging page loads.  Worse, each event recorded will have to be processed, tripling the database load.

Designed to Scale

You can relieve the pressure on the user by making your Pixel Tracking asynchronous.  Moving away from using your database as a queue is more complicated, but critical for scaling.  Finally, using Topics makes it easy to expand the types of processing tasks your platform supports.

Users are now completely insulated from scale and processing issues.

Dead Database Design

There is no database in the final design because it is no longer relevant to the users’ interactions with your services.  The performance is the same whether your database is at 0% or 100% load.  

The performance is the same if your database falls over and you have to switch to a hot standby or even restore from a backup.

With a bit of effort your SaaS could have a database fall over on the run up to Black Friday and recover without data loss or clients noticing. If you are using SNS/SQS on AWS the queue defaults are over 100,000 events! It may take a while to chew through the queues, but the data won’t disappear.

When your Pixel Tracking is causing your users headaches, going asynchronous is your Best Alternative to a Total Rewrite.

Putting Your Reputation on the Hook

This is a true story of how an inappropriate software design put a SaaS company’s reputation on the hook for their client’s bugs.  All names have been changed.

Once upon a time I consulted for Chicago Freight Brokers, a freight brokerage that was struggling with the performance of their in house systems.  CFB had an internal service, Express, that talked to 5 different transportation SaaS companies, added “special sauce” business value, and helped the brokers match trucks with loads.  Express was an MVP that was being crushed under the weight of its own success.

One of SaaS companies Express integrated with was ChicagoTMS.  ChicagoTMS, was a Transportation Management Service that built loads of freight and spoke to CFB’s accounting and insurance systems.  When a broker hit save on ChicagoTMS’s website, they would push the data into Express using a Webhook. One day I got asked to jump on a call, “ChicagoTMS is unbearably slow, it takes 30 seconds to save a load!  And they say it’s our fault! Can you talk to them?”

I jumped onto the call expecting that something had been lost in translation, after all, how could CFB’s scaling issues have any impact on ChicagoTMS’s web performance.

An hour later, I understood.  ChicagoTMS’s save operation looked like this:

The Webhook was called Synchronously as part of the save process, meaning users had to wait for the Webhook to complete successfully before they could continue their work.  CFB’s scaling problems were causing ChicagoTMS’s performance problems!

Fortunately, the culprit was a missing database index and the save time quickly dropped from 30 seconds to 2 seconds.  Unfortunately for ChicagoTMS, their design made them responsible for their client’s performance issues. ChicagoTMS’s reputation took a hit from CFB’s simple mistake.

A Better Design

Webhooks are convenient, but they are not a reliable or guaranteed delivery mechanism.  

A better design would have been for ChicagoTMS to make their Webhooks asynchronous.  This would have kept ChicagoTMS’s services fast and responsive, regardless of whether CFB made mistakes, did a lot of processing of the message, or was overwhelmed by a spike in load.

Asynchronous Webhooks would have given the users a much better experience and saved ChicagoTMS a lot of repetitional damage with their clients.

TAKE Your Reputation oFF the Hook!

Webhook performance will always be outside your control.  When you call a client synchronously you put your reputation in the client’s hands.  Your users won’t understand or care who exactly is responsible, if it’s your website, you get the blame.  

Go asynchronous and take your reputation off the hook!

I wouldn’t have done it this way.

I confess to having said, “I wouldn’t have done it this way.”  The phrase seems like a polite way of trashing the current system architecture while implying that you know the correct design.

You might get a snicker and feel smart and clever, but you’re creating problems for yourself and your project.

It puts the original programmers on the defensive.  

You may be politely trashing their design, but you’re still trashing their design.  Instead of listening to you describe your superior solution, the original programmers are going to be thinking of arguments defending their work.

The Business People Don’t Care

The managers and executives in the room didn’t care about how the original programmers designed the system.  They don’t care about how you would have designed the system. They want to know what you’re going to do about their business problems.

It leaves the listener questioning your intent.  

Are you changing the design because you don’t like it, or because it needs to be done?  You never want anyone wondering if you are proposing a refactor or rewrite because of style.

Your way might not be possible

I once used PostgreSQL as a noSql system because all the other AWS options had row size limits that were too small.  For years after new developers would tell me that I should have used different technologies. I would explain the size constraint, and more often than not, show that AWS still had constraints that would prevent us from using the technology.

Instead of trashing the original design, try these two approaches instead:

Talk about the business problem with the current design.

“The current design won’t scale to the levels we need.”  Maybe it won’t scale because it was a bad design, maybe it was a massively successful MVP.  Either way, you need to replace it to go forward.

Be positive

“The current design was a great way to get started.”  If the original design was an abject failure, no one would be asking you to rebuild or expand it.  Acknowledge that, whatever its shortcomings, the original design moved the business forward.

Acknowledge your ignorance

“I don’t know what the original requirements were, but the current design isn’t a good fit for our needs”.  It’s useful to know why past choices were made so that you don’t miss requirements, and people are much more likely to tell you if you’re the first to admit you don’t know.

“I wouldn’t have done it this way” is an old developer cliche.  The developers who inherited your work are probably saying it about you right now.

Making Link Tracking Scale – Part 2 Edge Caching

In Part 1 of Making Link Tracking Scale I showed how switching event recording from synchronous to asynchronous processing creates a superior, faster and more consistent user experience.  In Part 2, I will discuss how Link Tracking scaling issues are governed by Long Tails, and how to overcome the initial burst using edge caching and tools like Memcache and Redis.

The Long Tails of Link Tracking

When your client sends an email campaign, publishes new content your link tracker will experience a giant burst of activity, which will quickly decay like this:

To illustrate with some numbers, imagine an email blast that results in 100,000 link tracking events.  80% of those will occur in the first hour.  

In our original design from Part 1, that would 22 URL lookups, and 22 inserts per second:

For simplicity, pretend that inserts and selects produce similar db load.  Your system would need to support 44 events/s to avoid slowdowns and frustrating your clients.

The asynchronous model:

Reduces the load to 22 URL lookups, and a controllable number of inserts.  Again for simplicity let’s go with 8 inserts/s, for a total of 30 events/s.  That’s a 1/3 reduction in load!

But, your system is still looking up the Original URL 22 times/s.  That’s a lot of unnecessary db load.

Edge Caching The Original URL

The Original URL is static data that can be cached on the web server instead of loaded from the database for each event.  Instead, each server would retrieve the Original URL from the db once, store it in memory, and reuse it as needed.

This effectively drops the lookup rate from 22 events/s to 0 events/s, reducing the db load to 8 events/s, a 55% drop!  Combined with the asynchronous processing improvements from Part 1, that’s an 80% reduction in max database load.

Edge Caching on the servers works for a while, but as your clients expand the number of URLs you’ll need to keep track of won’t fit in server memory.  At that point you’ll need to add in tools like Memcached or Redis.  Like web servers, these tools are a lot cheaper than scaling your database.

Consistent Load on the Database

The great thing about this design is that you can keep the db load consistent, regardless of the incoming traffic.  Whether the load is 44 events/s or 100 events/s you control the rate of asynchronous processing. So long as you have room on your servers for an internal queue, or if you use an external queue like RabbitMQ or SQS you can delay processing the events.

Scaling questions become discussions about cost and how quickly your clients need to see results.

Conclusion

Caching static data is a great way to reduce database load.  You can use prebuilt libraries like Guava for Java, cacheout for Python, or dozens of others.  You can also leverage distributed cache systems like Memcached and Redis. While there’s no such thing as a free lunch, web servers and distributed caches are much much cheaper to scale than databases. 

You’ll save money and deliver a superior experience to your clients and their users!