Google-scale systems need multiple layers of load balancing and caching.
Load balancing: DNS-level (route to nearest data center), L4 (TCP connections), L7 (HTTP routing). Google uses Maglev, a custom L4 load balancer handling millions of connections per second.
Caching layers: CDN (static content close to users), reverse proxy cache (Varnish or custom), application cache (Memcached), database cache.
In your interview, mention specific caching strategies and explain cache invalidation. "Cache everything" is not a good answer. Explain what you cache, for how long, and how you handle stale data.