BCP · Business Continuity and Disaster Recovery
Business Continuity Plan & Disaster Recovery
What is BCP?
- Set recovery targets in numbers with RTO and RPO
- Having backups is not the same as being able to recover
- A plan that is never drilled does not work
Once a BCP is in place, the answer "Backups are running" becomes a commitment: "We restore the payment system within 4 hours, using last night's 11 p.m. data." Recovery no longer depends on one person's memory but on documents and access rights, and nobody argues on the day of an incident about what to bring up first. It is also a way to decide, in numbers and in advance, how much loss leadership can bear. The plan starts only when the heads of finance, sales, and customer service say for themselves how long their work can be down.
Recovery levels ultimately come down to money. A cold approach, which only stores backup media offsite, takes days to recover. A warm approach, with standby equipment already powered on, takes half a day. A hot approach, which fails over instantly through real-time replication, takes tens of minutes. Cutting RTO from 24 hours to 1 hour typically multiplies the cost several times over. That is why you don't apply the same tier to every system and instead rank them through a business impact analysis. For backups themselves, the baseline is the 3-2-1 rule: 3 copies, 2 media types, 1 offsite.
The boundaries between the three concepts also blur easily. Backup is the act of preserving data. DR is the IT work of bringing servers, networks, and accounts back. BCP also covers business-level alternatives, such as taking orders and paying salaries while systems are down. Scenarios like shipping on paper slips or running the call center from a backup office belong here. A company with DR alone gets its servers back but spends days with operations still stopped.
The most common slip is the claim, "It’s the cloud, so it’s handled." The provider’s availability promise covers infrastructure; wrongly deleted data and encrypted volumes are the customer’s responsibility. Ransomware locks the backup storage attached to the same network as production servers first, so without an isolated copy you have only added backup capacity. A documented RTO of 4 hours turning out to be twelve hours in a real drill is also common, and the cause is usually these three: accounts of departed employees, expired authentication certificates, and outdated contact lists.
Jaryong KooMarketing PhD and authority in AI, data analytics, and personal branding
John KimAuthor of B2B Marketing Design and CEO of Nanum Consulting| Category | Question | Example |
|---|---|---|
| RTO (Recovery Time Objective) | By when must we be running again? | Order system back online within 4 hours |
| RPO (Recovery Point Objective) | How far back must data be recovered | Up to 1 hour of data loss tolerated |
It evolved from disaster preparedness plans in industries where outages are fatal, such as finance and telecom. The 2012 ISO 22301 international standard then established a global benchmark for business continuity management systems. In Korea, industry-specific regulations and certifications also set related requirements.
Company A is a Korean auto parts maker with 240 employees and KRW 110 billion (≈ US$82,298,400) in annual revenue. Its MES and ERP, which run a two-shift line, sat in the same server room. A power outage halted them for eight hours, and the company paid KRW 36 million (≈ US$26,900) in late-delivery penalties. The head of management support ordered a recovery plan within three months. (Hypothetical example)
- The head of the information strategy team ran a business impact analysis with seven department heads. When the production management head insisted, "If the MES goes down, the line stops within two hours," the MES RTO dropped from 4 hours to 2.
- After a budget review, only the MES and ERP were upgraded to the warm tier, with standby servers kept running. Groupware and the HR system stayed cold, with a 48-hour RTO. That cut a KRW 180 million (≈ US$134,700) plan to KRW 62 million (≈ US$46,400) a year.
- While writing the recovery sequence, the IT operations manager found that three admin accounts were still registered under the names of employees who left two years earlier. All three were reissued, and the overnight emergency contact list was rebuilt to cover 31 people, including partner companies.
- To prepare for ransomware, full backups are taken weekly and stored in two places: a NAS disconnected from the network and an offsite vault. The backup owner and the deletion approver are separate people, so no one person can erase them.
- The mock recovery drill, held on a Saturday six months later, took 9 hours 40 minutes the first time. After the 12 issues it flagged were fixed, the MES came back in 1 hour 50 minutes in the second drill, and the ERP's data-loss window also shrank from 24 hours to 1 hour.